probe.c 80.4 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2
/*
B
Bjorn Helgaas 已提交
3
 * PCI detection and setup code
L
Linus Torvalds 已提交
4 5 6 7 8 9
 */

#include <linux/kernel.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/pci.h>
10
#include <linux/of_device.h>
11
#include <linux/of_pci.h>
12
#include <linux/pci_hotplug.h>
L
Linus Torvalds 已提交
13 14 15
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/cpumask.h>
S
Shaohua Li 已提交
16
#include <linux/pci-aspm.h>
17
#include <linux/aer.h>
18
#include <linux/acpi.h>
19
#include <linux/hypervisor.h>
20
#include <linux/irqdomain.h>
21
#include <linux/pm_runtime.h>
22
#include "pci.h"
L
Linus Torvalds 已提交
23 24 25 26

#define CARDBUS_LATENCY_TIMER	176	/* secondary latency timer */
#define CARDBUS_RESERVE_BUSNR	3

27
static struct resource busn_resource = {
Y
Yinghai Lu 已提交
28 29 30 31 32 33
	.name	= "PCI busn",
	.start	= 0,
	.end	= 255,
	.flags	= IORESOURCE_BUS,
};

L
Linus Torvalds 已提交
34 35 36 37
/* Ugh.  Need to stop exporting this to modules. */
LIST_HEAD(pci_root_buses);
EXPORT_SYMBOL(pci_root_buses);

38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
static LIST_HEAD(pci_domain_busn_res_list);

struct pci_domain_busn_res {
	struct list_head list;
	struct resource res;
	int domain_nr;
};

static struct resource *get_pci_domain_busn_res(int domain_nr)
{
	struct pci_domain_busn_res *r;

	list_for_each_entry(r, &pci_domain_busn_res_list, list)
		if (r->domain_nr == domain_nr)
			return &r->res;

	r = kzalloc(sizeof(*r), GFP_KERNEL);
	if (!r)
		return NULL;

	r->domain_nr = domain_nr;
	r->res.start = 0;
	r->res.end = 0xff;
	r->res.flags = IORESOURCE_BUS | IORESOURCE_PCI_FIXED;

	list_add_tail(&r->list, &pci_domain_busn_res_list);

	return &r->res;
}

68 69 70 71
static int find_anything(struct device *dev, void *data)
{
	return 1;
}
L
Linus Torvalds 已提交
72

Z
Zhang, Yanmin 已提交
73
/*
B
Bjorn Helgaas 已提交
74 75
 * Some device drivers need know if PCI is initiated.
 * Basically, we think PCI is not initiated when there
76
 * is no device to be found on the pci_bus_type.
Z
Zhang, Yanmin 已提交
77 78 79
 */
int no_pci_devices(void)
{
80 81
	struct device *dev;
	int no_devices;
Z
Zhang, Yanmin 已提交
82

83 84 85 86 87
	dev = bus_find_device(&pci_bus_type, NULL, NULL, find_anything);
	no_devices = (dev == NULL);
	put_device(dev);
	return no_devices;
}
Z
Zhang, Yanmin 已提交
88 89
EXPORT_SYMBOL(no_pci_devices);

L
Linus Torvalds 已提交
90 91 92
/*
 * PCI Bus Class
 */
93
static void release_pcibus_dev(struct device *dev)
L
Linus Torvalds 已提交
94
{
95
	struct pci_bus *pci_bus = to_pci_bus(dev);
L
Linus Torvalds 已提交
96

97
	put_device(pci_bus->bridge);
98
	pci_bus_remove_resources(pci_bus);
99
	pci_release_bus_of_node(pci_bus);
L
Linus Torvalds 已提交
100 101 102 103 104
	kfree(pci_bus);
}

static struct class pcibus_class = {
	.name		= "pci_bus",
105
	.dev_release	= &release_pcibus_dev,
106
	.dev_groups	= pcibus_groups,
L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114
};

static int __init pcibus_class_init(void)
{
	return class_register(&pcibus_class);
}
postcore_initcall(pcibus_class_init);

115
static u64 pci_size(u64 base, u64 maxbase, u64 mask)
L
Linus Torvalds 已提交
116
{
117
	u64 size = mask & maxbase;	/* Find the significant bits */
L
Linus Torvalds 已提交
118 119 120
	if (!size)
		return 0;

B
Bjorn Helgaas 已提交
121 122 123 124
	/*
	 * Get the lowest of them to find the decode size, and from that
	 * the extent.
	 */
L
Linus Torvalds 已提交
125 126
	size = (size & ~(size-1)) - 1;

B
Bjorn Helgaas 已提交
127 128 129 130
	/*
	 * base == maxbase can be valid only if the BAR has already been
	 * programmed with all 1s.
	 */
L
Linus Torvalds 已提交
131 132 133 134 135 136
	if (base == maxbase && ((base | size) & mask) != mask)
		return 0;

	return size;
}

137
static inline unsigned long decode_bar(struct pci_dev *dev, u32 bar)
138
{
139
	u32 mem_type;
140
	unsigned long flags;
141

142
	if ((bar & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
143 144 145
		flags = bar & ~PCI_BASE_ADDRESS_IO_MASK;
		flags |= IORESOURCE_IO;
		return flags;
146
	}
147

148 149 150 151
	flags = bar & ~PCI_BASE_ADDRESS_MEM_MASK;
	flags |= IORESOURCE_MEM;
	if (flags & PCI_BASE_ADDRESS_MEM_PREFETCH)
		flags |= IORESOURCE_PREFETCH;
152

153 154 155 156 157
	mem_type = bar & PCI_BASE_ADDRESS_MEM_TYPE_MASK;
	switch (mem_type) {
	case PCI_BASE_ADDRESS_MEM_TYPE_32:
		break;
	case PCI_BASE_ADDRESS_MEM_TYPE_1M:
158
		/* 1M mem BAR treated as 32-bit BAR */
159 160
		break;
	case PCI_BASE_ADDRESS_MEM_TYPE_64:
161 162
		flags |= IORESOURCE_MEM_64;
		break;
163
	default:
164
		/* mem unknown type treated as 32-bit BAR */
165 166
		break;
	}
167
	return flags;
168 169
}

170 171
#define PCI_COMMAND_DECODE_ENABLE	(PCI_COMMAND_MEMORY | PCI_COMMAND_IO)

Y
Yu Zhao 已提交
172
/**
B
Bjorn Helgaas 已提交
173
 * pci_read_base - Read a PCI BAR
Y
Yu Zhao 已提交
174 175 176 177 178 179
 * @dev: the PCI device
 * @type: type of the BAR
 * @res: resource buffer to be filled in
 * @pos: BAR position in the config space
 *
 * Returns 1 if the BAR is 64-bit, or 0 if 32-bit.
180
 */
Y
Yu Zhao 已提交
181
int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type,
R
Ryan Desfosses 已提交
182
		    struct resource *res, unsigned int pos)
183
{
184
	u32 l = 0, sz = 0, mask;
185
	u64 l64, sz64, mask64;
J
Jacob Pan 已提交
186
	u16 orig_cmd;
187
	struct pci_bus_region region, inverted_region;
188

189
	mask = type ? PCI_ROM_ADDRESS_MASK : ~0;
190

191
	/* No printks while decoding is disabled! */
J
Jacob Pan 已提交
192 193
	if (!dev->mmio_always_on) {
		pci_read_config_word(dev, PCI_COMMAND, &orig_cmd);
194 195 196 197
		if (orig_cmd & PCI_COMMAND_DECODE_ENABLE) {
			pci_write_config_word(dev, PCI_COMMAND,
				orig_cmd & ~PCI_COMMAND_DECODE_ENABLE);
		}
J
Jacob Pan 已提交
198 199
	}

200 201 202
	res->name = pci_name(dev);

	pci_read_config_dword(dev, pos, &l);
203
	pci_write_config_dword(dev, pos, l | mask);
204 205 206 207 208
	pci_read_config_dword(dev, pos, &sz);
	pci_write_config_dword(dev, pos, l);

	/*
	 * All bits set in sz means the device isn't working properly.
209 210 211
	 * If the BAR isn't implemented, all bits must be 0.  If it's a
	 * memory BAR or a ROM, bit 0 must be clear; if it's an io BAR, bit
	 * 1 must be clear.
212
	 */
213 214
	if (sz == 0xffffffff)
		sz = 0;
215 216 217 218 219 220 221 222 223

	/*
	 * I don't know how l can have all bits set.  Copied from old code.
	 * Maybe it fixes a bug on some ancient platform.
	 */
	if (l == 0xffffffff)
		l = 0;

	if (type == pci_bar_unknown) {
224 225 226
		res->flags = decode_bar(dev, l);
		res->flags |= IORESOURCE_SIZEALIGN;
		if (res->flags & IORESOURCE_IO) {
227 228 229
			l64 = l & PCI_BASE_ADDRESS_IO_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_IO_MASK;
			mask64 = PCI_BASE_ADDRESS_IO_MASK & (u32)IO_SPACE_LIMIT;
230
		} else {
231 232 233
			l64 = l & PCI_BASE_ADDRESS_MEM_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_MEM_MASK;
			mask64 = (u32)PCI_BASE_ADDRESS_MEM_MASK;
234 235
		}
	} else {
236 237
		if (l & PCI_ROM_ADDRESS_ENABLE)
			res->flags |= IORESOURCE_ROM_ENABLE;
238 239
		l64 = l & PCI_ROM_ADDRESS_MASK;
		sz64 = sz & PCI_ROM_ADDRESS_MASK;
240
		mask64 = PCI_ROM_ADDRESS_MASK;
241 242
	}

243
	if (res->flags & IORESOURCE_MEM_64) {
244 245 246 247 248 249 250
		pci_read_config_dword(dev, pos + 4, &l);
		pci_write_config_dword(dev, pos + 4, ~0);
		pci_read_config_dword(dev, pos + 4, &sz);
		pci_write_config_dword(dev, pos + 4, l);

		l64 |= ((u64)l << 32);
		sz64 |= ((u64)sz << 32);
251 252
		mask64 |= ((u64)~0 << 32);
	}
253

254 255
	if (!dev->mmio_always_on && (orig_cmd & PCI_COMMAND_DECODE_ENABLE))
		pci_write_config_word(dev, PCI_COMMAND, orig_cmd);
256

257 258
	if (!sz64)
		goto fail;
259

260
	sz64 = pci_size(l64, sz64, mask64);
261
	if (!sz64) {
262
		pci_info(dev, FW_BUG "reg 0x%x: invalid BAR (can't size)\n",
263
			 pos);
264
		goto fail;
265
	}
266 267

	if (res->flags & IORESOURCE_MEM_64) {
Y
Yinghai Lu 已提交
268 269
		if ((sizeof(pci_bus_addr_t) < 8 || sizeof(resource_size_t) < 8)
		    && sz64 > 0x100000000ULL) {
270 271 272
			res->flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
			res->start = 0;
			res->end = 0;
273
			pci_err(dev, "reg 0x%x: can't handle BAR larger than 4GB (size %#010llx)\n",
274
				pos, (unsigned long long)sz64);
275
			goto out;
276 277
		}

Y
Yinghai Lu 已提交
278
		if ((sizeof(pci_bus_addr_t) < 8) && l) {
279
			/* Above 32-bit boundary; try to reallocate */
280
			res->flags |= IORESOURCE_UNSET;
281 282
			res->start = 0;
			res->end = sz64;
283
			pci_info(dev, "reg 0x%x: can't handle BAR above 4GB (bus address %#010llx)\n",
284
				 pos, (unsigned long long)l64);
285
			goto out;
286 287 288
		}
	}

289 290 291
	region.start = l64;
	region.end = l64 + sz64;

292 293
	pcibios_bus_to_resource(dev->bus, res, &region);
	pcibios_resource_to_bus(dev->bus, &inverted_region, res);
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308

	/*
	 * If "A" is a BAR value (a bus address), "bus_to_resource(A)" is
	 * the corresponding resource address (the physical address used by
	 * the CPU.  Converting that resource address back to a bus address
	 * should yield the original BAR value:
	 *
	 *     resource_to_bus(bus_to_resource(A)) == A
	 *
	 * If it doesn't, CPU accesses to "bus_to_resource(A)" will not
	 * be claimed by the device.
	 */
	if (inverted_region.start != region.start) {
		res->flags |= IORESOURCE_UNSET;
		res->start = 0;
309
		res->end = region.end - region.start;
310
		pci_info(dev, "reg 0x%x: initial BAR value %#010llx invalid\n",
311
			 pos, (unsigned long long)region.start);
312
	}
313

314 315 316 317 318 319
	goto out;


fail:
	res->flags = 0;
out:
320
	if (res->flags)
321
		pci_printk(KERN_DEBUG, dev, "reg 0x%x: %pR\n", pos, res);
322

323
	return (res->flags & IORESOURCE_MEM_64) ? 1 : 0;
324 325
}

L
Linus Torvalds 已提交
326 327
static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
{
328
	unsigned int pos, reg;
329

330 331 332
	if (dev->non_compliant_bars)
		return;

333 334 335 336
	/* Per PCIe r4.0, sec 9.3.4.1.11, the VF BARs are all RO Zero */
	if (dev->is_virtfn)
		return;

337 338
	for (pos = 0; pos < howmany; pos++) {
		struct resource *res = &dev->resource[pos];
L
Linus Torvalds 已提交
339
		reg = PCI_BASE_ADDRESS_0 + (pos << 2);
340
		pos += __pci_read_base(dev, pci_bar_unknown, res, reg);
L
Linus Torvalds 已提交
341
	}
342

L
Linus Torvalds 已提交
343
	if (rom) {
344
		struct resource *res = &dev->resource[PCI_ROM_RESOURCE];
L
Linus Torvalds 已提交
345
		dev->rom_base_reg = rom;
346
		res->flags = IORESOURCE_MEM | IORESOURCE_PREFETCH |
347
				IORESOURCE_READONLY | IORESOURCE_SIZEALIGN;
348
		__pci_read_base(dev, pci_bar_mem32, res, rom);
L
Linus Torvalds 已提交
349 350 351
	}
}

B
Bill Pemberton 已提交
352
static void pci_read_bridge_io(struct pci_bus *child)
L
Linus Torvalds 已提交
353 354 355
{
	struct pci_dev *dev = child->self;
	u8 io_base_lo, io_limit_lo;
356
	unsigned long io_mask, io_granularity, base, limit;
357
	struct pci_bus_region region;
358 359 360 361 362 363 364 365 366
	struct resource *res;

	io_mask = PCI_IO_RANGE_MASK;
	io_granularity = 0x1000;
	if (dev->io_window_1k) {
		/* Support 1K I/O space granularity */
		io_mask = PCI_IO_1K_RANGE_MASK;
		io_granularity = 0x400;
	}
L
Linus Torvalds 已提交
367 368 369 370

	res = child->resource[0];
	pci_read_config_byte(dev, PCI_IO_BASE, &io_base_lo);
	pci_read_config_byte(dev, PCI_IO_LIMIT, &io_limit_lo);
371 372
	base = (io_base_lo & io_mask) << 8;
	limit = (io_limit_lo & io_mask) << 8;
L
Linus Torvalds 已提交
373 374 375

	if ((io_base_lo & PCI_IO_RANGE_TYPE_MASK) == PCI_IO_RANGE_TYPE_32) {
		u16 io_base_hi, io_limit_hi;
376

L
Linus Torvalds 已提交
377 378
		pci_read_config_word(dev, PCI_IO_BASE_UPPER16, &io_base_hi);
		pci_read_config_word(dev, PCI_IO_LIMIT_UPPER16, &io_limit_hi);
379 380
		base |= ((unsigned long) io_base_hi << 16);
		limit |= ((unsigned long) io_limit_hi << 16);
L
Linus Torvalds 已提交
381 382
	}

383
	if (base <= limit) {
L
Linus Torvalds 已提交
384
		res->flags = (io_base_lo & PCI_IO_RANGE_TYPE_MASK) | IORESOURCE_IO;
385
		region.start = base;
386
		region.end = limit + io_granularity - 1;
387
		pcibios_bus_to_resource(dev->bus, res, &region);
388
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
389
	}
390 391
}

B
Bill Pemberton 已提交
392
static void pci_read_bridge_mmio(struct pci_bus *child)
393 394 395 396
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
397
	struct pci_bus_region region;
398
	struct resource *res;
L
Linus Torvalds 已提交
399 400 401 402

	res = child->resource[1];
	pci_read_config_word(dev, PCI_MEMORY_BASE, &mem_base_lo);
	pci_read_config_word(dev, PCI_MEMORY_LIMIT, &mem_limit_lo);
403 404
	base = ((unsigned long) mem_base_lo & PCI_MEMORY_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_MEMORY_RANGE_MASK) << 16;
405
	if (base <= limit) {
L
Linus Torvalds 已提交
406
		res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM;
407 408
		region.start = base;
		region.end = limit + 0xfffff;
409
		pcibios_bus_to_resource(dev->bus, res, &region);
410
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
411
	}
412 413
}

B
Bill Pemberton 已提交
414
static void pci_read_bridge_mmio_pref(struct pci_bus *child)
415 416 417
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
418
	u64 base64, limit64;
Y
Yinghai Lu 已提交
419
	pci_bus_addr_t base, limit;
420
	struct pci_bus_region region;
421
	struct resource *res;
L
Linus Torvalds 已提交
422 423 424 425

	res = child->resource[2];
	pci_read_config_word(dev, PCI_PREF_MEMORY_BASE, &mem_base_lo);
	pci_read_config_word(dev, PCI_PREF_MEMORY_LIMIT, &mem_limit_lo);
426 427
	base64 = (mem_base_lo & PCI_PREF_RANGE_MASK) << 16;
	limit64 = (mem_limit_lo & PCI_PREF_RANGE_MASK) << 16;
L
Linus Torvalds 已提交
428 429 430

	if ((mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) == PCI_PREF_RANGE_TYPE_64) {
		u32 mem_base_hi, mem_limit_hi;
431

L
Linus Torvalds 已提交
432 433 434 435 436 437 438 439 440
		pci_read_config_dword(dev, PCI_PREF_BASE_UPPER32, &mem_base_hi);
		pci_read_config_dword(dev, PCI_PREF_LIMIT_UPPER32, &mem_limit_hi);

		/*
		 * Some bridges set the base > limit by default, and some
		 * (broken) BIOSes do not initialize them.  If we find
		 * this, just assume they are not being used.
		 */
		if (mem_base_hi <= mem_limit_hi) {
441 442
			base64 |= (u64) mem_base_hi << 32;
			limit64 |= (u64) mem_limit_hi << 32;
L
Linus Torvalds 已提交
443 444
		}
	}
445

Y
Yinghai Lu 已提交
446 447
	base = (pci_bus_addr_t) base64;
	limit = (pci_bus_addr_t) limit64;
448 449

	if (base != base64) {
450
		pci_err(dev, "can't handle bridge window above 4GB (bus address %#010llx)\n",
451 452 453 454
			(unsigned long long) base64);
		return;
	}

455
	if (base <= limit) {
456 457 458 459
		res->flags = (mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) |
					 IORESOURCE_MEM | IORESOURCE_PREFETCH;
		if (res->flags & PCI_PREF_RANGE_TYPE_64)
			res->flags |= IORESOURCE_MEM_64;
460 461
		region.start = base;
		region.end = limit + 0xfffff;
462
		pcibios_bus_to_resource(dev->bus, res, &region);
463
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
464 465 466
	}
}

B
Bill Pemberton 已提交
467
void pci_read_bridge_bases(struct pci_bus *child)
468 469
{
	struct pci_dev *dev = child->self;
470
	struct resource *res;
471 472 473 474 475
	int i;

	if (pci_is_root_bus(child))	/* It's a host bus, nothing to read */
		return;

476
	pci_info(dev, "PCI bridge to %pR%s\n",
477
		 &child->busn_res,
478 479
		 dev->transparent ? " (subtractive decode)" : "");

480 481 482 483
	pci_bus_remove_resources(child);
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++)
		child->resource[i] = &dev->resource[PCI_BRIDGE_RESOURCES+i];

484 485 486
	pci_read_bridge_io(child);
	pci_read_bridge_mmio(child);
	pci_read_bridge_mmio_pref(child);
487 488

	if (dev->transparent) {
489
		pci_bus_for_each_resource(child->parent, res, i) {
490
			if (res && res->flags) {
491 492
				pci_bus_add_resource(child, res,
						     PCI_SUBTRACTIVE_DECODE);
493
				pci_printk(KERN_DEBUG, dev,
494
					   "  bridge window %pR (subtractive decode)\n",
495 496
					   res);
			}
497 498
		}
	}
499 500
}

501
static struct pci_bus *pci_alloc_bus(struct pci_bus *parent)
L
Linus Torvalds 已提交
502 503 504
{
	struct pci_bus *b;

505
	b = kzalloc(sizeof(*b), GFP_KERNEL);
506 507 508 509 510 511 512 513 514 515
	if (!b)
		return NULL;

	INIT_LIST_HEAD(&b->node);
	INIT_LIST_HEAD(&b->children);
	INIT_LIST_HEAD(&b->devices);
	INIT_LIST_HEAD(&b->slots);
	INIT_LIST_HEAD(&b->resources);
	b->max_bus_speed = PCI_SPEED_UNKNOWN;
	b->cur_bus_speed = PCI_SPEED_UNKNOWN;
516 517 518 519
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	if (parent)
		b->domain_nr = parent->domain_nr;
#endif
L
Linus Torvalds 已提交
520 521 522
	return b;
}

523
static void devm_pci_release_host_bridge_dev(struct device *dev)
524 525 526 527 528
{
	struct pci_host_bridge *bridge = to_pci_host_bridge(dev);

	if (bridge->release_fn)
		bridge->release_fn(bridge);
529 530

	pci_free_resource_list(&bridge->windows);
531
}
532

533 534 535
static void pci_release_host_bridge_dev(struct device *dev)
{
	devm_pci_release_host_bridge_dev(dev);
536
	kfree(to_pci_host_bridge(dev));
537 538
}

539
struct pci_host_bridge *pci_alloc_host_bridge(size_t priv)
540 541 542
{
	struct pci_host_bridge *bridge;

543
	bridge = kzalloc(sizeof(*bridge) + priv, GFP_KERNEL);
544 545
	if (!bridge)
		return NULL;
546

547
	INIT_LIST_HEAD(&bridge->windows);
548
	bridge->dev.release = pci_release_host_bridge_dev;
549

550 551 552 553 554 555 556
	/*
	 * We assume we can manage these PCIe features.  Some systems may
	 * reserve these for use by the platform itself, e.g., an ACPI BIOS
	 * may implement its own AER handling and use _OSC to prevent the
	 * OS from interfering.
	 */
	bridge->native_aer = 1;
557
	bridge->native_pcie_hotplug = 1;
558
	bridge->native_shpc_hotplug = 1;
559
	bridge->native_pme = 1;
560
	bridge->native_ltr = 1;
561

562 563
	return bridge;
}
564
EXPORT_SYMBOL(pci_alloc_host_bridge);
565

566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
struct pci_host_bridge *devm_pci_alloc_host_bridge(struct device *dev,
						   size_t priv)
{
	struct pci_host_bridge *bridge;

	bridge = devm_kzalloc(dev, sizeof(*bridge) + priv, GFP_KERNEL);
	if (!bridge)
		return NULL;

	INIT_LIST_HEAD(&bridge->windows);
	bridge->dev.release = devm_pci_release_host_bridge_dev;

	return bridge;
}
EXPORT_SYMBOL(devm_pci_alloc_host_bridge);

582 583 584 585 586 587 588 589
void pci_free_host_bridge(struct pci_host_bridge *bridge)
{
	pci_free_resource_list(&bridge->windows);

	kfree(bridge);
}
EXPORT_SYMBOL(pci_free_host_bridge);

590
static const unsigned char pcix_bus_speed[] = {
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCI_SPEED_66MHz_PCIX,		/* 1 */
	PCI_SPEED_100MHz_PCIX,		/* 2 */
	PCI_SPEED_133MHz_PCIX,		/* 3 */
	PCI_SPEED_UNKNOWN,		/* 4 */
	PCI_SPEED_66MHz_PCIX_ECC,	/* 5 */
	PCI_SPEED_100MHz_PCIX_ECC,	/* 6 */
	PCI_SPEED_133MHz_PCIX_ECC,	/* 7 */
	PCI_SPEED_UNKNOWN,		/* 8 */
	PCI_SPEED_66MHz_PCIX_266,	/* 9 */
	PCI_SPEED_100MHz_PCIX_266,	/* A */
	PCI_SPEED_133MHz_PCIX_266,	/* B */
	PCI_SPEED_UNKNOWN,		/* C */
	PCI_SPEED_66MHz_PCIX_533,	/* D */
	PCI_SPEED_100MHz_PCIX_533,	/* E */
	PCI_SPEED_133MHz_PCIX_533	/* F */
};

609
const unsigned char pcie_link_speed[] = {
610 611 612
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCIE_SPEED_2_5GT,		/* 1 */
	PCIE_SPEED_5_0GT,		/* 2 */
613
	PCIE_SPEED_8_0GT,		/* 3 */
614
	PCIE_SPEED_16_0GT,		/* 4 */
615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
	PCI_SPEED_UNKNOWN,		/* 5 */
	PCI_SPEED_UNKNOWN,		/* 6 */
	PCI_SPEED_UNKNOWN,		/* 7 */
	PCI_SPEED_UNKNOWN,		/* 8 */
	PCI_SPEED_UNKNOWN,		/* 9 */
	PCI_SPEED_UNKNOWN,		/* A */
	PCI_SPEED_UNKNOWN,		/* B */
	PCI_SPEED_UNKNOWN,		/* C */
	PCI_SPEED_UNKNOWN,		/* D */
	PCI_SPEED_UNKNOWN,		/* E */
	PCI_SPEED_UNKNOWN		/* F */
};

void pcie_update_link_speed(struct pci_bus *bus, u16 linksta)
{
630
	bus->cur_bus_speed = pcie_link_speed[linksta & PCI_EXP_LNKSTA_CLS];
631 632 633
}
EXPORT_SYMBOL_GPL(pcie_update_link_speed);

634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653
static unsigned char agp_speeds[] = {
	AGP_UNKNOWN,
	AGP_1X,
	AGP_2X,
	AGP_4X,
	AGP_8X
};

static enum pci_bus_speed agp_speed(int agp3, int agpstat)
{
	int index = 0;

	if (agpstat & 4)
		index = 3;
	else if (agpstat & 2)
		index = 2;
	else if (agpstat & 1)
		index = 1;
	else
		goto out;
654

655 656 657 658 659 660 661 662 663 664
	if (agp3) {
		index += 2;
		if (index == 5)
			index = 0;
	}

 out:
	return agp_speeds[index];
}

665 666 667 668 669
static void pci_set_bus_speed(struct pci_bus *bus)
{
	struct pci_dev *bridge = bus->self;
	int pos;

670 671 672 673 674 675 676 677 678 679 680 681 682
	pos = pci_find_capability(bridge, PCI_CAP_ID_AGP);
	if (!pos)
		pos = pci_find_capability(bridge, PCI_CAP_ID_AGP3);
	if (pos) {
		u32 agpstat, agpcmd;

		pci_read_config_dword(bridge, pos + PCI_AGP_STATUS, &agpstat);
		bus->max_bus_speed = agp_speed(agpstat & 8, agpstat & 7);

		pci_read_config_dword(bridge, pos + PCI_AGP_COMMAND, &agpcmd);
		bus->cur_bus_speed = agp_speed(agpstat & 8, agpcmd & 7);
	}

683 684 685 686 687
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (pos) {
		u16 status;
		enum pci_bus_speed max;

688 689 690 691
		pci_read_config_word(bridge, pos + PCI_X_BRIDGE_SSTATUS,
				     &status);

		if (status & PCI_X_SSTATUS_533MHZ) {
692
			max = PCI_SPEED_133MHz_PCIX_533;
693
		} else if (status & PCI_X_SSTATUS_266MHZ) {
694
			max = PCI_SPEED_133MHz_PCIX_266;
695
		} else if (status & PCI_X_SSTATUS_133MHZ) {
R
Ryan Desfosses 已提交
696
			if ((status & PCI_X_SSTATUS_VERS) == PCI_X_SSTATUS_V2)
697
				max = PCI_SPEED_133MHz_PCIX_ECC;
R
Ryan Desfosses 已提交
698
			else
699 700 701 702 703 704
				max = PCI_SPEED_133MHz_PCIX;
		} else {
			max = PCI_SPEED_66MHz_PCIX;
		}

		bus->max_bus_speed = max;
705 706
		bus->cur_bus_speed = pcix_bus_speed[
			(status & PCI_X_SSTATUS_FREQ) >> 6];
707 708 709 710

		return;
	}

711
	if (pci_is_pcie(bridge)) {
712 713 714
		u32 linkcap;
		u16 linksta;

715
		pcie_capability_read_dword(bridge, PCI_EXP_LNKCAP, &linkcap);
716
		bus->max_bus_speed = pcie_link_speed[linkcap & PCI_EXP_LNKCAP_SLS];
717

718
		pcie_capability_read_word(bridge, PCI_EXP_LNKSTA, &linksta);
719 720 721 722
		pcie_update_link_speed(bus, linksta);
	}
}

723 724
static struct irq_domain *pci_host_bridge_msi_domain(struct pci_bus *bus)
{
725 726
	struct irq_domain *d;

727 728 729 730
	/*
	 * Any firmware interface that can resolve the msi_domain
	 * should be called from here.
	 */
731
	d = pci_host_bridge_of_msi_domain(bus);
732 733
	if (!d)
		d = pci_host_bridge_acpi_msi_domain(bus);
734

735 736 737 738 739 740 741 742 743 744 745 746 747 748
#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
	/*
	 * If no IRQ domain was found via the OF tree, try looking it up
	 * directly through the fwnode_handle.
	 */
	if (!d) {
		struct fwnode_handle *fwnode = pci_root_bus_fwnode(bus);

		if (fwnode)
			d = irq_find_matching_fwnode(fwnode,
						     DOMAIN_BUS_PCI_MSI);
	}
#endif

749
	return d;
750 751 752 753 754
}

static void pci_set_bus_msi_domain(struct pci_bus *bus)
{
	struct irq_domain *d;
755
	struct pci_bus *b;
756 757

	/*
758 759 760
	 * The bus can be a root bus, a subordinate bus, or a virtual bus
	 * created by an SR-IOV device.  Walk up to the first bridge device
	 * found or derive the domain from the host bridge.
761
	 */
762 763 764 765 766 767 768
	for (b = bus, d = NULL; !d && !pci_is_root_bus(b); b = b->parent) {
		if (b->self)
			d = dev_get_msi_domain(&b->self->dev);
	}

	if (!d)
		d = pci_host_bridge_msi_domain(b);
769 770 771 772

	dev_set_msi_domain(&bus->dev, d);
}

773
static int pci_register_host_bridge(struct pci_host_bridge *bridge)
774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790
{
	struct device *parent = bridge->dev.parent;
	struct resource_entry *window, *n;
	struct pci_bus *bus, *b;
	resource_size_t offset;
	LIST_HEAD(resources);
	struct resource *res;
	char addr[64], *fmt;
	const char *name;
	int err;

	bus = pci_alloc_bus(NULL);
	if (!bus)
		return -ENOMEM;

	bridge->bus = bus;

B
Bjorn Helgaas 已提交
791
	/* Temporarily move resources off the list */
792 793 794 795 796 797 798 799 800 801 802
	list_splice_init(&bridge->windows, &resources);
	bus->sysdata = bridge->sysdata;
	bus->msi = bridge->msi;
	bus->ops = bridge->ops;
	bus->number = bus->busn_res.start = bridge->busnr;
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	bus->domain_nr = pci_bus_find_domain_nr(bus, parent);
#endif

	b = pci_find_bus(pci_domain_nr(bus), bridge->busnr);
	if (b) {
B
Bjorn Helgaas 已提交
803
		/* Ignore it if we already got here via a different bridge */
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888
		dev_dbg(&b->dev, "bus already known\n");
		err = -EEXIST;
		goto free;
	}

	dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(bus),
		     bridge->busnr);

	err = pcibios_root_bridge_prepare(bridge);
	if (err)
		goto free;

	err = device_register(&bridge->dev);
	if (err)
		put_device(&bridge->dev);

	bus->bridge = get_device(&bridge->dev);
	device_enable_async_suspend(bus->bridge);
	pci_set_bus_of_node(bus);
	pci_set_bus_msi_domain(bus);

	if (!parent)
		set_dev_node(bus->bridge, pcibus_to_node(bus));

	bus->dev.class = &pcibus_class;
	bus->dev.parent = bus->bridge;

	dev_set_name(&bus->dev, "%04x:%02x", pci_domain_nr(bus), bus->number);
	name = dev_name(&bus->dev);

	err = device_register(&bus->dev);
	if (err)
		goto unregister;

	pcibios_add_bus(bus);

	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(bus);

	if (parent)
		dev_info(parent, "PCI host bridge to bus %s\n", name);
	else
		pr_info("PCI host bridge to bus %s\n", name);

	/* Add initial resources to the bus */
	resource_list_for_each_entry_safe(window, n, &resources) {
		list_move_tail(&window->node, &bridge->windows);
		offset = window->offset;
		res = window->res;

		if (res->flags & IORESOURCE_BUS)
			pci_bus_insert_busn_res(bus, bus->number, res->end);
		else
			pci_bus_add_resource(bus, res, 0);

		if (offset) {
			if (resource_type(res) == IORESOURCE_IO)
				fmt = " (bus address [%#06llx-%#06llx])";
			else
				fmt = " (bus address [%#010llx-%#010llx])";

			snprintf(addr, sizeof(addr), fmt,
				 (unsigned long long)(res->start - offset),
				 (unsigned long long)(res->end - offset));
		} else
			addr[0] = '\0';

		dev_info(&bus->dev, "root bus resource %pR%s\n", res, addr);
	}

	down_write(&pci_bus_sem);
	list_add_tail(&bus->node, &pci_root_buses);
	up_write(&pci_bus_sem);

	return 0;

unregister:
	put_device(&bridge->dev);
	device_unregister(&bridge->dev);

free:
	kfree(bus);
	return err;
}

889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
static bool pci_bridge_child_ext_cfg_accessible(struct pci_dev *bridge)
{
	int pos;
	u32 status;

	/*
	 * If extended config space isn't accessible on a bridge's primary
	 * bus, we certainly can't access it on the secondary bus.
	 */
	if (bridge->bus->bus_flags & PCI_BUS_FLAGS_NO_EXTCFG)
		return false;

	/*
	 * PCIe Root Ports and switch ports are PCIe on both sides, so if
	 * extended config space is accessible on the primary, it's also
	 * accessible on the secondary.
	 */
	if (pci_is_pcie(bridge) &&
	    (pci_pcie_type(bridge) == PCI_EXP_TYPE_ROOT_PORT ||
	     pci_pcie_type(bridge) == PCI_EXP_TYPE_UPSTREAM ||
	     pci_pcie_type(bridge) == PCI_EXP_TYPE_DOWNSTREAM))
		return true;

	/*
	 * For the other bridge types:
	 *   - PCI-to-PCI bridges
	 *   - PCIe-to-PCI/PCI-X forward bridges
	 *   - PCI/PCI-X-to-PCIe reverse bridges
	 * extended config space on the secondary side is only accessible
	 * if the bridge supports PCI-X Mode 2.
	 */
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (!pos)
		return false;

	pci_read_config_dword(bridge, pos + PCI_X_STATUS, &status);
	return status & (PCI_X_STATUS_266MHZ | PCI_X_STATUS_533MHZ);
}

928 929
static struct pci_bus *pci_alloc_child_bus(struct pci_bus *parent,
					   struct pci_dev *bridge, int busnr)
L
Linus Torvalds 已提交
930 931 932
{
	struct pci_bus *child;
	int i;
933
	int ret;
L
Linus Torvalds 已提交
934

B
Bjorn Helgaas 已提交
935
	/* Allocate a new bus and inherit stuff from the parent */
936
	child = pci_alloc_bus(parent);
L
Linus Torvalds 已提交
937 938 939 940 941
	if (!child)
		return NULL;

	child->parent = parent;
	child->ops = parent->ops;
942
	child->msi = parent->msi;
L
Linus Torvalds 已提交
943
	child->sysdata = parent->sysdata;
944
	child->bus_flags = parent->bus_flags;
L
Linus Torvalds 已提交
945

B
Bjorn Helgaas 已提交
946 947 948
	/*
	 * Initialize some portions of the bus device, but don't register
	 * it now as the parent is not properly set up yet.
949 950
	 */
	child->dev.class = &pcibus_class;
951
	dev_set_name(&child->dev, "%04x:%02x", pci_domain_nr(child), busnr);
L
Linus Torvalds 已提交
952

B
Bjorn Helgaas 已提交
953
	/* Set up the primary, secondary and subordinate bus numbers */
954 955 956
	child->number = child->busn_res.start = busnr;
	child->primary = parent->busn_res.start;
	child->busn_res.end = 0xff;
L
Linus Torvalds 已提交
957

958 959 960 961
	if (!bridge) {
		child->dev.parent = parent->bridge;
		goto add_dev;
	}
962 963 964

	child->self = bridge;
	child->bridge = get_device(&bridge->dev);
965
	child->dev.parent = child->bridge;
966
	pci_set_bus_of_node(child);
967 968
	pci_set_bus_speed(child);

969 970 971 972 973 974 975 976 977 978
	/*
	 * Check whether extended config space is accessible on the child
	 * bus.  Note that we currently assume it is always accessible on
	 * the root bus.
	 */
	if (!pci_bridge_child_ext_cfg_accessible(bridge)) {
		child->bus_flags |= PCI_BUS_FLAGS_NO_EXTCFG;
		pci_info(child, "extended config space not accessible\n");
	}

B
Bjorn Helgaas 已提交
979
	/* Set up default resource pointers and names */
980
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++) {
L
Linus Torvalds 已提交
981 982 983 984 985
		child->resource[i] = &bridge->resource[PCI_BRIDGE_RESOURCES+i];
		child->resource[i]->name = child->name;
	}
	bridge->subordinate = child;

986
add_dev:
987
	pci_set_bus_msi_domain(child);
988 989 990
	ret = device_register(&child->dev);
	WARN_ON(ret < 0);

991 992
	pcibios_add_bus(child);

993 994 995 996 997 998
	if (child->ops->add_bus) {
		ret = child->ops->add_bus(child);
		if (WARN_ON(ret < 0))
			dev_err(&child->dev, "failed to add bus: %d\n", ret);
	}

999 1000 1001
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(child);

L
Linus Torvalds 已提交
1002 1003 1004
	return child;
}

R
Ryan Desfosses 已提交
1005 1006
struct pci_bus *pci_add_new_bus(struct pci_bus *parent, struct pci_dev *dev,
				int busnr)
L
Linus Torvalds 已提交
1007 1008 1009 1010
{
	struct pci_bus *child;

	child = pci_alloc_child_bus(parent, dev, busnr);
1011
	if (child) {
1012
		down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
1013
		list_add_tail(&child->node, &parent->children);
1014
		up_write(&pci_bus_sem);
1015
	}
L
Linus Torvalds 已提交
1016 1017
	return child;
}
1018
EXPORT_SYMBOL(pci_add_new_bus);
L
Linus Torvalds 已提交
1019

1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
static void pci_enable_crs(struct pci_dev *pdev)
{
	u16 root_cap = 0;

	/* Enable CRS Software Visibility if supported */
	pcie_capability_read_word(pdev, PCI_EXP_RTCAP, &root_cap);
	if (root_cap & PCI_EXP_RTCAP_CRSVIS)
		pcie_capability_set_word(pdev, PCI_EXP_RTCTL,
					 PCI_EXP_RTCTL_CRSSVE);
}

1031 1032 1033
static unsigned int pci_scan_child_bus_extend(struct pci_bus *bus,
					      unsigned int available_buses);

L
Linus Torvalds 已提交
1034
/*
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
 * pci_scan_bridge_extend() - Scan buses behind a bridge
 * @bus: Parent bus the bridge is on
 * @dev: Bridge itself
 * @max: Starting subordinate number of buses behind this bridge
 * @available_buses: Total number of buses available for this bridge and
 *		     the devices below. After the minimal bus space has
 *		     been allocated the remaining buses will be
 *		     distributed equally between hotplug-capable bridges.
 * @pass: Either %0 (scan already configured bridges) or %1 (scan bridges
 *        that need to be reconfigured.
 *
L
Linus Torvalds 已提交
1046 1047 1048 1049 1050 1051 1052 1053
 * If it's a bridge, configure it and scan the bus behind it.
 * For CardBus bridges, we don't scan behind as the devices will
 * be handled by the bridge driver itself.
 *
 * We need to process bridges in two passes -- first we scan those
 * already configured by the BIOS and after we are done with all of
 * them, we proceed to assigning numbers to the remaining buses in
 * order to avoid overlaps between old and new bus numbers.
1054 1055
 *
 * Return: New subordinate number covering all buses behind this bridge.
L
Linus Torvalds 已提交
1056
 */
1057 1058 1059
static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
				  int max, unsigned int available_buses,
				  int pass)
L
Linus Torvalds 已提交
1060 1061 1062
{
	struct pci_bus *child;
	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
1063
	u32 buses, i, j = 0;
L
Linus Torvalds 已提交
1064
	u16 bctl;
1065
	u8 primary, secondary, subordinate;
1066
	int broken = 0;
L
Linus Torvalds 已提交
1067

1068 1069 1070 1071 1072 1073
	/*
	 * Make sure the bridge is powered on to be able to access config
	 * space of devices below it.
	 */
	pm_runtime_get_sync(&dev->dev);

L
Linus Torvalds 已提交
1074
	pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
1075 1076 1077
	primary = buses & 0xFF;
	secondary = (buses >> 8) & 0xFF;
	subordinate = (buses >> 16) & 0xFF;
L
Linus Torvalds 已提交
1078

1079
	pci_dbg(dev, "scanning [bus %02x-%02x] behind bridge, pass %d\n",
1080
		secondary, subordinate, pass);
L
Linus Torvalds 已提交
1081

1082
	if (!primary && (primary != bus->number) && secondary && subordinate) {
1083
		pci_warn(dev, "Primary bus is hard wired to 0\n");
1084 1085 1086
		primary = bus->number;
	}

1087 1088
	/* Check if setup is sensible at all */
	if (!pass &&
1089
	    (primary != bus->number || secondary <= bus->number ||
1090
	     secondary > subordinate)) {
1091
		pci_info(dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
1092
			 secondary, subordinate);
1093 1094 1095
		broken = 1;
	}

B
Bjorn Helgaas 已提交
1096 1097 1098 1099
	/*
	 * Disable Master-Abort Mode during probing to avoid reporting of
	 * bus errors in some architectures.
	 */
L
Linus Torvalds 已提交
1100 1101 1102 1103
	pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &bctl);
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL,
			      bctl & ~PCI_BRIDGE_CTL_MASTER_ABORT);

1104 1105
	pci_enable_crs(dev);

1106 1107 1108
	if ((secondary || subordinate) && !pcibios_assign_all_busses() &&
	    !is_cardbus && !broken) {
		unsigned int cmax;
B
Bjorn Helgaas 已提交
1109

L
Linus Torvalds 已提交
1110
		/*
B
Bjorn Helgaas 已提交
1111 1112
		 * Bus already configured by firmware, process it in the
		 * first pass and just note the configuration.
L
Linus Torvalds 已提交
1113 1114
		 */
		if (pass)
1115
			goto out;
L
Linus Torvalds 已提交
1116 1117

		/*
B
Bjorn Helgaas 已提交
1118 1119 1120 1121
		 * The bus might already exist for two reasons: Either we
		 * are rescanning the bus or the bus is reachable through
		 * more than one bridge. The second case can happen with
		 * the i450NX chipset.
L
Linus Torvalds 已提交
1122
		 */
1123
		child = pci_find_bus(pci_domain_nr(bus), secondary);
A
Alex Chiang 已提交
1124
		if (!child) {
1125
			child = pci_add_new_bus(bus, dev, secondary);
A
Alex Chiang 已提交
1126 1127
			if (!child)
				goto out;
1128
			child->primary = primary;
Y
Yinghai Lu 已提交
1129
			pci_bus_insert_busn_res(child, secondary, subordinate);
A
Alex Chiang 已提交
1130
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
1131 1132 1133
		}

		cmax = pci_scan_child_bus(child);
1134
		if (cmax > subordinate)
1135
			pci_warn(dev, "bridge has subordinate %02x but max busn %02x\n",
1136
				 subordinate, cmax);
B
Bjorn Helgaas 已提交
1137 1138

		/* Subordinate should equal child->busn_res.end */
1139 1140
		if (subordinate > max)
			max = subordinate;
L
Linus Torvalds 已提交
1141
	} else {
B
Bjorn Helgaas 已提交
1142

L
Linus Torvalds 已提交
1143 1144 1145 1146
		/*
		 * We need to assign a number to this bus which we always
		 * do in the second pass.
		 */
1147
		if (!pass) {
1148
			if (pcibios_assign_all_busses() || broken || is_cardbus)
B
Bjorn Helgaas 已提交
1149 1150 1151 1152 1153 1154 1155 1156 1157

				/*
				 * Temporarily disable forwarding of the
				 * configuration cycles on all bridges in
				 * this bus segment to avoid possible
				 * conflicts in the second pass between two
				 * bridges programmed with overlapping bus
				 * ranges.
				 */
1158 1159
				pci_write_config_dword(dev, PCI_PRIMARY_BUS,
						       buses & ~0xffffff);
1160
			goto out;
1161
		}
L
Linus Torvalds 已提交
1162 1163 1164 1165

		/* Clear errors */
		pci_write_config_word(dev, PCI_STATUS, 0xffff);

B
Bjorn Helgaas 已提交
1166 1167 1168 1169 1170
		/*
		 * Prevent assigning a bus number that already exists.
		 * This can happen when a bridge is hot-plugged, so in this
		 * case we only re-scan this bus.
		 */
1171 1172
		child = pci_find_bus(pci_domain_nr(bus), max+1);
		if (!child) {
1173
			child = pci_add_new_bus(bus, dev, max+1);
1174 1175
			if (!child)
				goto out;
1176 1177
			pci_bus_insert_busn_res(child, max+1,
						bus->busn_res.end);
1178
		}
1179
		max++;
1180 1181 1182
		if (available_buses)
			available_buses--;

L
Linus Torvalds 已提交
1183 1184
		buses = (buses & 0xff000000)
		      | ((unsigned int)(child->primary)     <<  0)
1185 1186
		      | ((unsigned int)(child->busn_res.start)   <<  8)
		      | ((unsigned int)(child->busn_res.end) << 16);
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192 1193 1194 1195

		/*
		 * yenta.c forces a secondary latency timer of 176.
		 * Copy that behaviour here.
		 */
		if (is_cardbus) {
			buses &= ~0xff000000;
			buses |= CARDBUS_LATENCY_TIMER << 24;
		}
1196

B
Bjorn Helgaas 已提交
1197
		/* We need to blast all three values with a single write */
L
Linus Torvalds 已提交
1198 1199 1200
		pci_write_config_dword(dev, PCI_PRIMARY_BUS, buses);

		if (!is_cardbus) {
1201
			child->bridge_ctl = bctl;
1202
			max = pci_scan_child_bus_extend(child, available_buses);
L
Linus Torvalds 已提交
1203
		} else {
B
Bjorn Helgaas 已提交
1204

L
Linus Torvalds 已提交
1205
			/*
B
Bjorn Helgaas 已提交
1206 1207 1208
			 * For CardBus bridges, we leave 4 bus numbers as
			 * cards with a PCI-to-PCI bridge can be inserted
			 * later.
L
Linus Torvalds 已提交
1209
			 */
R
Ryan Desfosses 已提交
1210
			for (i = 0; i < CARDBUS_RESERVE_BUSNR; i++) {
1211
				struct pci_bus *parent = bus;
1212 1213 1214
				if (pci_find_bus(pci_domain_nr(bus),
							max+i+1))
					break;
1215 1216
				while (parent->parent) {
					if ((!pcibios_assign_all_busses()) &&
1217 1218
					    (parent->busn_res.end > max) &&
					    (parent->busn_res.end <= max+i)) {
1219 1220 1221 1222 1223
						j = 1;
					}
					parent = parent->parent;
				}
				if (j) {
B
Bjorn Helgaas 已提交
1224

1225
					/*
B
Bjorn Helgaas 已提交
1226 1227 1228
					 * Often, there are two CardBus
					 * bridges -- try to leave one
					 * valid bus number for each one.
1229 1230 1231 1232 1233
					 */
					i /= 2;
					break;
				}
			}
1234
			max += i;
L
Linus Torvalds 已提交
1235
		}
B
Bjorn Helgaas 已提交
1236 1237

		/* Set subordinate bus number to its real value */
Y
Yinghai Lu 已提交
1238
		pci_bus_update_busn_res_end(child, max);
L
Linus Torvalds 已提交
1239 1240 1241
		pci_write_config_byte(dev, PCI_SUBORDINATE_BUS, max);
	}

1242 1243 1244
	sprintf(child->name,
		(is_cardbus ? "PCI CardBus %04x:%02x" : "PCI Bus %04x:%02x"),
		pci_domain_nr(bus), child->number);
L
Linus Torvalds 已提交
1245

1246
	/* Check that all devices are accessible */
1247
	while (bus->parent) {
1248 1249
		if ((child->busn_res.end > bus->busn_res.end) ||
		    (child->number > bus->busn_res.end) ||
1250
		    (child->number < bus->number) ||
1251
		    (child->busn_res.end < bus->number)) {
1252 1253 1254
			dev_info(&dev->dev, "devices behind bridge are unusable because %pR cannot be assigned for them\n",
				 &child->busn_res);
			break;
1255 1256 1257 1258
		}
		bus = bus->parent;
	}

1259 1260 1261
out:
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);

1262 1263
	pm_runtime_put(&dev->dev);

L
Linus Torvalds 已提交
1264 1265
	return max;
}
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282

/*
 * pci_scan_bridge() - Scan buses behind a bridge
 * @bus: Parent bus the bridge is on
 * @dev: Bridge itself
 * @max: Starting subordinate number of buses behind this bridge
 * @pass: Either %0 (scan already configured bridges) or %1 (scan bridges
 *        that need to be reconfigured.
 *
 * If it's a bridge, configure it and scan the bus behind it.
 * For CardBus bridges, we don't scan behind as the devices will
 * be handled by the bridge driver itself.
 *
 * We need to process bridges in two passes -- first we scan those
 * already configured by the BIOS and after we are done with all of
 * them, we proceed to assigning numbers to the remaining buses in
 * order to avoid overlaps between old and new bus numbers.
1283 1284
 *
 * Return: New subordinate number covering all buses behind this bridge.
1285 1286 1287 1288 1289
 */
int pci_scan_bridge(struct pci_bus *bus, struct pci_dev *dev, int max, int pass)
{
	return pci_scan_bridge_extend(bus, dev, max, 0, pass);
}
1290
EXPORT_SYMBOL(pci_scan_bridge);
L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299

/*
 * Read interrupt line and base address registers.
 * The architecture-dependent code can tweak these, of course.
 */
static void pci_read_irq(struct pci_dev *dev)
{
	unsigned char irq;

1300 1301 1302 1303 1304 1305 1306
	/* VFs are not allowed to use INTx, so skip the config reads */
	if (dev->is_virtfn) {
		dev->pin = 0;
		dev->irq = 0;
		return;
	}

L
Linus Torvalds 已提交
1307
	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &irq);
1308
	dev->pin = irq;
L
Linus Torvalds 已提交
1309 1310 1311 1312 1313
	if (irq)
		pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
	dev->irq = irq;
}

1314
void set_pcie_port_type(struct pci_dev *pdev)
Y
Yu Zhao 已提交
1315 1316 1317
{
	int pos;
	u16 reg16;
1318 1319
	int type;
	struct pci_dev *parent;
Y
Yu Zhao 已提交
1320 1321 1322 1323

	pos = pci_find_capability(pdev, PCI_CAP_ID_EXP);
	if (!pos)
		return;
1324

1325
	pdev->pcie_cap = pos;
Y
Yu Zhao 已提交
1326
	pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, &reg16);
1327
	pdev->pcie_flags_reg = reg16;
1328 1329
	pci_read_config_word(pdev, pos + PCI_EXP_DEVCAP, &reg16);
	pdev->pcie_mpss = reg16 & PCI_EXP_DEVCAP_PAYLOAD;
1330 1331

	/*
1332 1333 1334 1335
	 * A Root Port or a PCI-to-PCIe bridge is always the upstream end
	 * of a Link.  No PCIe component has two Links.  Two Links are
	 * connected by a Switch that has a Port on each Link and internal
	 * logic to connect the two Ports.
1336 1337
	 */
	type = pci_pcie_type(pdev);
1338 1339
	if (type == PCI_EXP_TYPE_ROOT_PORT ||
	    type == PCI_EXP_TYPE_PCIE_BRIDGE)
1340 1341 1342 1343
		pdev->has_secondary_link = 1;
	else if (type == PCI_EXP_TYPE_UPSTREAM ||
		 type == PCI_EXP_TYPE_DOWNSTREAM) {
		parent = pci_upstream_bridge(pdev);
1344 1345 1346 1347 1348 1349

		/*
		 * Usually there's an upstream device (Root Port or Switch
		 * Downstream Port), but we can't assume one exists.
		 */
		if (parent && !parent->has_secondary_link)
1350 1351
			pdev->has_secondary_link = 1;
	}
Y
Yu Zhao 已提交
1352 1353
}

1354
void set_pcie_hotplug_bridge(struct pci_dev *pdev)
1355 1356 1357
{
	u32 reg32;

1358
	pcie_capability_read_dword(pdev, PCI_EXP_SLTCAP, &reg32);
1359 1360 1361 1362
	if (reg32 & PCI_EXP_SLTCAP_HPC)
		pdev->is_hotplug_bridge = 1;
}

1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
static void set_pcie_thunderbolt(struct pci_dev *dev)
{
	int vsec = 0;
	u32 header;

	while ((vsec = pci_find_next_ext_capability(dev, vsec,
						    PCI_EXT_CAP_ID_VNDR))) {
		pci_read_config_dword(dev, vsec + PCI_VNDR_HEADER, &header);

		/* Is the device part of a Thunderbolt controller? */
		if (dev->vendor == PCI_VENDOR_ID_INTEL &&
		    PCI_VNDR_HEADER_ID(header) == PCI_VSEC_ID_INTEL_TBT) {
			dev->is_thunderbolt = 1;
			return;
		}
	}
}

1381
/**
B
Bjorn Helgaas 已提交
1382
 * pci_ext_cfg_is_aliased - Is ext config space just an alias of std config?
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
 * @dev: PCI device
 *
 * PCI Express to PCI/PCI-X Bridge Specification, rev 1.0, 4.1.4 says that
 * when forwarding a type1 configuration request the bridge must check that
 * the extended register address field is zero.  The bridge is not permitted
 * to forward the transactions and must handle it as an Unsupported Request.
 * Some bridges do not follow this rule and simply drop the extended register
 * bits, resulting in the standard config space being aliased, every 256
 * bytes across the entire configuration space.  Test for this condition by
 * comparing the first dword of each potential alias to the vendor/device ID.
 * Known offenders:
 *   ASM1083/1085 PCIe-to-PCI Reversible Bridge (1b21:1080, rev 01 & 03)
 *   AMD/ATI SBx00 PCI to PCI Bridge (1002:4384, rev 40)
 */
static bool pci_ext_cfg_is_aliased(struct pci_dev *dev)
{
#ifdef CONFIG_PCI_QUIRKS
	int pos;
	u32 header, tmp;

	pci_read_config_dword(dev, PCI_VENDOR_ID, &header);

	for (pos = PCI_CFG_SPACE_SIZE;
	     pos < PCI_CFG_SPACE_EXP_SIZE; pos += PCI_CFG_SPACE_SIZE) {
		if (pci_read_config_dword(dev, pos, &tmp) != PCIBIOS_SUCCESSFUL
		    || header != tmp)
			return false;
	}

	return true;
#else
	return false;
#endif
}

1418
/**
B
Bjorn Helgaas 已提交
1419
 * pci_cfg_space_size - Get the configuration space size of the PCI device
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
 * @dev: PCI device
 *
 * Regular PCI devices have 256 bytes, but PCI-X 2 and PCI Express devices
 * have 4096 bytes.  Even if the device is capable, that doesn't mean we can
 * access it.  Maybe we don't have a way to generate extended config space
 * accesses, or the device is behind a reverse Express bridge.  So we try
 * reading the dword at 0x100 which must either be 0 or a valid extended
 * capability header.
 */
static int pci_cfg_space_size_ext(struct pci_dev *dev)
{
	u32 status;
	int pos = PCI_CFG_SPACE_SIZE;

	if (pci_read_config_dword(dev, pos, &status) != PCIBIOS_SUCCESSFUL)
1435
		return PCI_CFG_SPACE_SIZE;
1436
	if (status == 0xffffffff || pci_ext_cfg_is_aliased(dev))
1437
		return PCI_CFG_SPACE_SIZE;
1438 1439 1440 1441 1442 1443 1444 1445 1446 1447

	return PCI_CFG_SPACE_EXP_SIZE;
}

int pci_cfg_space_size(struct pci_dev *dev)
{
	int pos;
	u32 status;
	u16 class;

1448 1449 1450
	if (dev->bus->bus_flags & PCI_BUS_FLAGS_NO_EXTCFG)
		return PCI_CFG_SPACE_SIZE;

1451 1452 1453 1454
	class = dev->class >> 8;
	if (class == PCI_CLASS_BRIDGE_HOST)
		return pci_cfg_space_size_ext(dev);

1455 1456
	if (pci_is_pcie(dev))
		return pci_cfg_space_size_ext(dev);
1457

1458 1459 1460
	pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
	if (!pos)
		return PCI_CFG_SPACE_SIZE;
1461

1462 1463 1464
	pci_read_config_dword(dev, pos + PCI_X_STATUS, &status);
	if (status & (PCI_X_STATUS_266MHZ | PCI_X_STATUS_533MHZ))
		return pci_cfg_space_size_ext(dev);
1465 1466 1467 1468

	return PCI_CFG_SPACE_SIZE;
}

1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
static u32 pci_class(struct pci_dev *dev)
{
	u32 class;

#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn)
		return dev->physfn->sriov->class;
#endif
	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);
	return class;
}

static void pci_subsystem_ids(struct pci_dev *dev, u16 *vendor, u16 *device)
{
#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn) {
		*vendor = dev->physfn->sriov->subsystem_vendor;
		*device = dev->physfn->sriov->subsystem_device;
		return;
	}
#endif
	pci_read_config_word(dev, PCI_SUBSYSTEM_VENDOR_ID, vendor);
	pci_read_config_word(dev, PCI_SUBSYSTEM_ID, device);
}

static u8 pci_hdr_type(struct pci_dev *dev)
{
	u8 hdr_type;

#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn)
		return dev->physfn->sriov->hdr_type;
#endif
	pci_read_config_byte(dev, PCI_HEADER_TYPE, &hdr_type);
	return hdr_type;
}

1506
#define LEGACY_IO_RESOURCE	(IORESOURCE_IO | IORESOURCE_PCI_FIXED)
1507

1508
static void pci_msi_setup_pci_dev(struct pci_dev *dev)
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523
{
	/*
	 * Disable the MSI hardware to avoid screaming interrupts
	 * during boot.  This is the power on reset default so
	 * usually this should be a noop.
	 */
	dev->msi_cap = pci_find_capability(dev, PCI_CAP_ID_MSI);
	if (dev->msi_cap)
		pci_msi_set_enable(dev, 0);

	dev->msix_cap = pci_find_capability(dev, PCI_CAP_ID_MSIX);
	if (dev->msix_cap)
		pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_ENABLE, 0);
}

1524
/**
B
Bjorn Helgaas 已提交
1525
 * pci_intx_mask_broken - Test PCI_COMMAND_INTX_DISABLE writability
1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
 * @dev: PCI device
 *
 * Test whether PCI_COMMAND_INTX_DISABLE is writable for @dev.  Check this
 * at enumeration-time to avoid modifying PCI_COMMAND at run-time.
 */
static int pci_intx_mask_broken(struct pci_dev *dev)
{
	u16 orig, toggle, new;

	pci_read_config_word(dev, PCI_COMMAND, &orig);
	toggle = orig ^ PCI_COMMAND_INTX_DISABLE;
	pci_write_config_word(dev, PCI_COMMAND, toggle);
	pci_read_config_word(dev, PCI_COMMAND, &new);

	pci_write_config_word(dev, PCI_COMMAND, orig);

	/*
	 * PCI_COMMAND_INTX_DISABLE was reserved and read-only prior to PCI
	 * r2.3, so strictly speaking, a device is not *broken* if it's not
	 * writable.  But we'll live with the misnomer for now.
	 */
	if (new != toggle)
		return 1;
	return 0;
}

L
Linus Torvalds 已提交
1552
/**
B
Bjorn Helgaas 已提交
1553
 * pci_setup_device - Fill in class and map information of a device
L
Linus Torvalds 已提交
1554 1555
 * @dev: the device structure to fill
 *
1556
 * Initialize the device structure with information about the device's
B
Bjorn Helgaas 已提交
1557
 * vendor,class,memory and IO-space addresses, IRQ lines etc.
L
Linus Torvalds 已提交
1558
 * Called at initialisation of the PCI subsystem and by CardBus services.
Y
Yu Zhao 已提交
1559 1560
 * Returns 0 on success and negative if unknown type of device (not normal,
 * bridge or CardBus).
L
Linus Torvalds 已提交
1561
 */
Y
Yu Zhao 已提交
1562
int pci_setup_device(struct pci_dev *dev)
L
Linus Torvalds 已提交
1563 1564
{
	u32 class;
1565
	u16 cmd;
Y
Yu Zhao 已提交
1566
	u8 hdr_type;
1567
	int pos = 0;
1568 1569
	struct pci_bus_region region;
	struct resource *res;
Y
Yu Zhao 已提交
1570

1571
	hdr_type = pci_hdr_type(dev);
Y
Yu Zhao 已提交
1572 1573 1574 1575 1576 1577 1578 1579 1580

	dev->sysdata = dev->bus->sysdata;
	dev->dev.parent = dev->bus->bridge;
	dev->dev.bus = &pci_bus_type;
	dev->hdr_type = hdr_type & 0x7f;
	dev->multifunction = !!(hdr_type & 0x80);
	dev->error_state = pci_channel_io_normal;
	set_pcie_port_type(dev);

1581
	pci_dev_assign_slot(dev);
B
Bjorn Helgaas 已提交
1582 1583 1584 1585 1586

	/*
	 * Assume 32-bit PCI; let 64-bit PCI cards (which are far rarer)
	 * set this higher, assuming the system even supports it.
	 */
Y
Yu Zhao 已提交
1587
	dev->dma_mask = 0xffffffff;
L
Linus Torvalds 已提交
1588

1589 1590 1591
	dev_set_name(&dev->dev, "%04x:%02x:%02x.%d", pci_domain_nr(dev->bus),
		     dev->bus->number, PCI_SLOT(dev->devfn),
		     PCI_FUNC(dev->devfn));
L
Linus Torvalds 已提交
1592

1593 1594
	class = pci_class(dev);

A
Auke Kok 已提交
1595
	dev->revision = class & 0xff;
Y
Yinghai Lu 已提交
1596
	dev->class = class >> 8;		    /* upper 3 bytes */
L
Linus Torvalds 已提交
1597

1598
	pci_printk(KERN_DEBUG, dev, "[%04x:%04x] type %02x class %#08x\n",
Y
Yinghai Lu 已提交
1599
		   dev->vendor, dev->device, dev->hdr_type, dev->class);
L
Linus Torvalds 已提交
1600

B
Bjorn Helgaas 已提交
1601
	/* Need to have dev->class ready */
1602 1603
	dev->cfg_size = pci_cfg_space_size(dev);

B
Bjorn Helgaas 已提交
1604
	/* Need to have dev->cfg_size ready */
1605 1606
	set_pcie_thunderbolt(dev);

L
Linus Torvalds 已提交
1607
	/* "Unknown power state" */
1608
	dev->current_state = PCI_UNKNOWN;
L
Linus Torvalds 已提交
1609 1610 1611

	/* Early fixups, before probing the BARs */
	pci_fixup_device(pci_fixup_early, dev);
B
Bjorn Helgaas 已提交
1612 1613

	/* Device class may be changed after fixup */
1614
	class = dev->class >> 8;
L
Linus Torvalds 已提交
1615

1616 1617 1618
	if (dev->non_compliant_bars) {
		pci_read_config_word(dev, PCI_COMMAND, &cmd);
		if (cmd & (PCI_COMMAND_IO | PCI_COMMAND_MEMORY)) {
1619
			pci_info(dev, "device has non-compliant BARs; disabling IO/MEM decoding\n");
1620 1621 1622 1623 1624 1625
			cmd &= ~PCI_COMMAND_IO;
			cmd &= ~PCI_COMMAND_MEMORY;
			pci_write_config_word(dev, PCI_COMMAND, cmd);
		}
	}

1626 1627
	dev->broken_intx_masking = pci_intx_mask_broken(dev);

L
Linus Torvalds 已提交
1628 1629 1630 1631 1632 1633
	switch (dev->hdr_type) {		    /* header type */
	case PCI_HEADER_TYPE_NORMAL:		    /* standard header */
		if (class == PCI_CLASS_BRIDGE_PCI)
			goto bad;
		pci_read_irq(dev);
		pci_read_bases(dev, 6, PCI_ROM_ADDRESS);
1634 1635

		pci_subsystem_ids(dev, &dev->subsystem_vendor, &dev->subsystem_device);
1636 1637

		/*
1638 1639 1640 1641
		 * Do the ugly legacy mode stuff here rather than broken chip
		 * quirk code. Legacy mode ATA controllers have fixed
		 * addresses. These are not always echoed in BAR0-3, and
		 * BAR0-3 in a few cases contain junk!
1642 1643 1644 1645 1646
		 */
		if (class == PCI_CLASS_STORAGE_IDE) {
			u8 progif;
			pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
			if ((progif & 1) == 0) {
1647 1648 1649 1650
				region.start = 0x1F0;
				region.end = 0x1F7;
				res = &dev->resource[0];
				res->flags = LEGACY_IO_RESOURCE;
1651
				pcibios_bus_to_resource(dev->bus, res, &region);
1652
				pci_info(dev, "legacy IDE quirk: reg 0x10: %pR\n",
1653
					 res);
1654 1655 1656 1657
				region.start = 0x3F6;
				region.end = 0x3F6;
				res = &dev->resource[1];
				res->flags = LEGACY_IO_RESOURCE;
1658
				pcibios_bus_to_resource(dev->bus, res, &region);
1659
				pci_info(dev, "legacy IDE quirk: reg 0x14: %pR\n",
1660
					 res);
1661 1662
			}
			if ((progif & 4) == 0) {
1663 1664 1665 1666
				region.start = 0x170;
				region.end = 0x177;
				res = &dev->resource[2];
				res->flags = LEGACY_IO_RESOURCE;
1667
				pcibios_bus_to_resource(dev->bus, res, &region);
1668
				pci_info(dev, "legacy IDE quirk: reg 0x18: %pR\n",
1669
					 res);
1670 1671 1672 1673
				region.start = 0x376;
				region.end = 0x376;
				res = &dev->resource[3];
				res->flags = LEGACY_IO_RESOURCE;
1674
				pcibios_bus_to_resource(dev->bus, res, &region);
1675
				pci_info(dev, "legacy IDE quirk: reg 0x1c: %pR\n",
1676
					 res);
1677 1678
			}
		}
L
Linus Torvalds 已提交
1679 1680 1681 1682 1683
		break;

	case PCI_HEADER_TYPE_BRIDGE:		    /* bridge header */
		if (class != PCI_CLASS_BRIDGE_PCI)
			goto bad;
B
Bjorn Helgaas 已提交
1684 1685 1686 1687 1688 1689

		/*
		 * The PCI-to-PCI bridge spec requires that subtractive
		 * decoding (i.e. transparent) bridge must have programming
		 * interface code of 0x01.
		 */
1690
		pci_read_irq(dev);
L
Linus Torvalds 已提交
1691 1692
		dev->transparent = ((dev->class & 0xff) == 1);
		pci_read_bases(dev, 2, PCI_ROM_ADDRESS1);
1693
		set_pcie_hotplug_bridge(dev);
1694 1695 1696 1697 1698
		pos = pci_find_capability(dev, PCI_CAP_ID_SSVID);
		if (pos) {
			pci_read_config_word(dev, pos + PCI_SSVID_VENDOR_ID, &dev->subsystem_vendor);
			pci_read_config_word(dev, pos + PCI_SSVID_DEVICE_ID, &dev->subsystem_device);
		}
L
Linus Torvalds 已提交
1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		break;

	case PCI_HEADER_TYPE_CARDBUS:		    /* CardBus bridge header */
		if (class != PCI_CLASS_BRIDGE_CARDBUS)
			goto bad;
		pci_read_irq(dev);
		pci_read_bases(dev, 1, 0);
		pci_read_config_word(dev, PCI_CB_SUBSYSTEM_VENDOR_ID, &dev->subsystem_vendor);
		pci_read_config_word(dev, PCI_CB_SUBSYSTEM_ID, &dev->subsystem_device);
		break;

	default:				    /* unknown header */
1711
		pci_err(dev, "unknown header type %02x, ignoring device\n",
1712
			dev->hdr_type);
Y
Yu Zhao 已提交
1713
		return -EIO;
L
Linus Torvalds 已提交
1714 1715

	bad:
1716
		pci_err(dev, "ignoring class %#08x (doesn't match header type %02x)\n",
1717
			dev->class, dev->hdr_type);
1718
		dev->class = PCI_CLASS_NOT_DEFINED << 8;
L
Linus Torvalds 已提交
1719 1720 1721 1722 1723 1724
	}

	/* We found a fine healthy device, go go go... */
	return 0;
}

1725 1726 1727
static void pci_configure_mps(struct pci_dev *dev)
{
	struct pci_dev *bridge = pci_upstream_bridge(dev);
1728
	int mps, p_mps, rc;
1729 1730 1731 1732

	if (!pci_is_pcie(dev) || !bridge || !pci_is_pcie(bridge))
		return;

1733 1734 1735 1736
	/* MPS and MRRS fields are of type 'RsvdP' for VFs, short-circuit out */
	if (dev->is_virtfn)
		return;

1737 1738 1739 1740 1741 1742 1743
	mps = pcie_get_mps(dev);
	p_mps = pcie_get_mps(bridge);

	if (mps == p_mps)
		return;

	if (pcie_bus_config == PCIE_BUS_TUNE_OFF) {
1744
		pci_warn(dev, "Max Payload Size %d, but upstream %s set to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
1745 1746 1747
			 mps, pci_name(bridge), p_mps);
		return;
	}
1748 1749 1750 1751 1752 1753 1754 1755 1756 1757

	/*
	 * Fancier MPS configuration is done later by
	 * pcie_bus_configure_settings()
	 */
	if (pcie_bus_config != PCIE_BUS_DEFAULT)
		return;

	rc = pcie_set_mps(dev, p_mps);
	if (rc) {
1758
		pci_warn(dev, "can't set Max Payload Size to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
1759 1760 1761 1762
			 p_mps);
		return;
	}

1763
	pci_info(dev, "Max Payload Size set to %d (was %d, max %d)\n",
1764
		 p_mps, mps, 128 << dev->pcie_mpss);
1765 1766
}

1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778
static struct hpp_type0 pci_default_type0 = {
	.revision = 1,
	.cache_line_size = 8,
	.latency_timer = 0x40,
	.enable_serr = 0,
	.enable_perr = 0,
};

static void program_hpp_type0(struct pci_dev *dev, struct hpp_type0 *hpp)
{
	u16 pci_cmd, pci_bctl;

1779
	if (!hpp)
1780 1781 1782
		hpp = &pci_default_type0;

	if (hpp->revision > 1) {
1783
		pci_warn(dev, "PCI settings rev %d not supported; using defaults\n",
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811
			 hpp->revision);
		hpp = &pci_default_type0;
	}

	pci_write_config_byte(dev, PCI_CACHE_LINE_SIZE, hpp->cache_line_size);
	pci_write_config_byte(dev, PCI_LATENCY_TIMER, hpp->latency_timer);
	pci_read_config_word(dev, PCI_COMMAND, &pci_cmd);
	if (hpp->enable_serr)
		pci_cmd |= PCI_COMMAND_SERR;
	if (hpp->enable_perr)
		pci_cmd |= PCI_COMMAND_PARITY;
	pci_write_config_word(dev, PCI_COMMAND, pci_cmd);

	/* Program bridge control value */
	if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI) {
		pci_write_config_byte(dev, PCI_SEC_LATENCY_TIMER,
				      hpp->latency_timer);
		pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &pci_bctl);
		if (hpp->enable_serr)
			pci_bctl |= PCI_BRIDGE_CTL_SERR;
		if (hpp->enable_perr)
			pci_bctl |= PCI_BRIDGE_CTL_PARITY;
		pci_write_config_word(dev, PCI_BRIDGE_CONTROL, pci_bctl);
	}
}

static void program_hpp_type1(struct pci_dev *dev, struct hpp_type1 *hpp)
{
1812 1813 1814 1815 1816 1817 1818 1819 1820
	int pos;

	if (!hpp)
		return;

	pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
	if (!pos)
		return;

1821
	pci_warn(dev, "PCI-X settings not supported\n");
1822 1823
}

1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
static bool pcie_root_rcb_set(struct pci_dev *dev)
{
	struct pci_dev *rp = pcie_find_root_port(dev);
	u16 lnkctl;

	if (!rp)
		return false;

	pcie_capability_read_word(rp, PCI_EXP_LNKCTL, &lnkctl);
	if (lnkctl & PCI_EXP_LNKCTL_RCB)
		return true;

	return false;
}

1839 1840 1841 1842 1843 1844 1845 1846
static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp)
{
	int pos;
	u32 reg32;

	if (!hpp)
		return;

1847 1848 1849
	if (!pci_is_pcie(dev))
		return;

1850
	if (hpp->revision > 1) {
1851
		pci_warn(dev, "PCIe settings rev %d not supported\n",
1852 1853 1854 1855
			 hpp->revision);
		return;
	}

1856 1857 1858 1859 1860 1861 1862 1863 1864 1865
	/*
	 * Don't allow _HPX to change MPS or MRRS settings.  We manage
	 * those to make sure they're consistent with the rest of the
	 * platform.
	 */
	hpp->pci_exp_devctl_and |= PCI_EXP_DEVCTL_PAYLOAD |
				    PCI_EXP_DEVCTL_READRQ;
	hpp->pci_exp_devctl_or &= ~(PCI_EXP_DEVCTL_PAYLOAD |
				    PCI_EXP_DEVCTL_READRQ);

1866 1867 1868 1869 1870
	/* Initialize Device Control Register */
	pcie_capability_clear_and_set_word(dev, PCI_EXP_DEVCTL,
			~hpp->pci_exp_devctl_and, hpp->pci_exp_devctl_or);

	/* Initialize Link Control Register */
1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881
	if (pcie_cap_has_lnkctl(dev)) {

		/*
		 * If the Root Port supports Read Completion Boundary of
		 * 128, set RCB to 128.  Otherwise, clear it.
		 */
		hpp->pci_exp_lnkctl_and |= PCI_EXP_LNKCTL_RCB;
		hpp->pci_exp_lnkctl_or &= ~PCI_EXP_LNKCTL_RCB;
		if (pcie_root_rcb_set(dev))
			hpp->pci_exp_lnkctl_or |= PCI_EXP_LNKCTL_RCB;

1882 1883
		pcie_capability_clear_and_set_word(dev, PCI_EXP_LNKCTL,
			~hpp->pci_exp_lnkctl_and, hpp->pci_exp_lnkctl_or);
1884
	}
1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908

	/* Find Advanced Error Reporting Enhanced Capability */
	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
	if (!pos)
		return;

	/* Initialize Uncorrectable Error Mask Register */
	pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, &reg32);
	reg32 = (reg32 & hpp->unc_err_mask_and) | hpp->unc_err_mask_or;
	pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, reg32);

	/* Initialize Uncorrectable Error Severity Register */
	pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, &reg32);
	reg32 = (reg32 & hpp->unc_err_sever_and) | hpp->unc_err_sever_or;
	pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, reg32);

	/* Initialize Correctable Error Mask Register */
	pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, &reg32);
	reg32 = (reg32 & hpp->cor_err_mask_and) | hpp->cor_err_mask_or;
	pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg32);

	/* Initialize Advanced Error Capabilities and Control Register */
	pci_read_config_dword(dev, pos + PCI_ERR_CAP, &reg32);
	reg32 = (reg32 & hpp->adv_err_cap_and) | hpp->adv_err_cap_or;
B
Bjorn Helgaas 已提交
1909

1910 1911 1912 1913 1914
	/* Don't enable ECRC generation or checking if unsupported */
	if (!(reg32 & PCI_ERR_CAP_ECRC_GENC))
		reg32 &= ~PCI_ERR_CAP_ECRC_GENE;
	if (!(reg32 & PCI_ERR_CAP_ECRC_CHKC))
		reg32 &= ~PCI_ERR_CAP_ECRC_CHKE;
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
	pci_write_config_dword(dev, pos + PCI_ERR_CAP, reg32);

	/*
	 * FIXME: The following two registers are not supported yet.
	 *
	 *   o Secondary Uncorrectable Error Severity Register
	 *   o Secondary Uncorrectable Error Mask Register
	 */
}

1925
int pci_configure_extended_tags(struct pci_dev *dev, void *ign)
1926
{
1927 1928 1929
	struct pci_host_bridge *host;
	u32 cap;
	u16 ctl;
1930 1931 1932
	int ret;

	if (!pci_is_pcie(dev))
1933
		return 0;
1934

1935
	ret = pcie_capability_read_dword(dev, PCI_EXP_DEVCAP, &cap);
1936
	if (ret)
1937 1938 1939 1940
		return 0;

	if (!(cap & PCI_EXP_DEVCAP_EXT_TAG))
		return 0;
1941

1942 1943 1944 1945 1946 1947 1948
	ret = pcie_capability_read_word(dev, PCI_EXP_DEVCTL, &ctl);
	if (ret)
		return 0;

	host = pci_find_host_bridge(dev->bus);
	if (!host)
		return 0;
1949

1950 1951 1952 1953 1954 1955
	/*
	 * If some device in the hierarchy doesn't handle Extended Tags
	 * correctly, make sure they're disabled.
	 */
	if (host->no_ext_tags) {
		if (ctl & PCI_EXP_DEVCTL_EXT_TAG) {
1956
			pci_info(dev, "disabling Extended Tags\n");
1957 1958 1959 1960 1961 1962 1963
			pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
						   PCI_EXP_DEVCTL_EXT_TAG);
		}
		return 0;
	}

	if (!(ctl & PCI_EXP_DEVCTL_EXT_TAG)) {
1964
		pci_info(dev, "enabling Extended Tags\n");
1965 1966
		pcie_capability_set_word(dev, PCI_EXP_DEVCTL,
					 PCI_EXP_DEVCTL_EXT_TAG);
1967 1968
	}
	return 0;
1969 1970
}

1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
/**
 * pcie_relaxed_ordering_enabled - Probe for PCIe relaxed ordering enable
 * @dev: PCI device to query
 *
 * Returns true if the device has enabled relaxed ordering attribute.
 */
bool pcie_relaxed_ordering_enabled(struct pci_dev *dev)
{
	u16 v;

	pcie_capability_read_word(dev, PCI_EXP_DEVCTL, &v);

	return !!(v & PCI_EXP_DEVCTL_RELAX_EN);
}
EXPORT_SYMBOL(pcie_relaxed_ordering_enabled);

static void pci_configure_relaxed_ordering(struct pci_dev *dev)
{
	struct pci_dev *root;

	/* PCI_EXP_DEVICE_RELAX_EN is RsvdP in VFs */
	if (dev->is_virtfn)
		return;

	if (!pcie_relaxed_ordering_enabled(dev))
		return;

	/*
	 * For now, we only deal with Relaxed Ordering issues with Root
	 * Ports. Peer-to-Peer DMA is another can of worms.
	 */
	root = pci_find_pcie_root_port(dev);
	if (!root)
		return;

	if (root->dev_flags & PCI_DEV_FLAGS_NO_RELAXED_ORDERING) {
		pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
					   PCI_EXP_DEVCTL_RELAX_EN);
2009
		pci_info(dev, "Relaxed Ordering disabled because the Root Port didn't support it\n");
2010 2011 2012
	}
}

2013 2014 2015
static void pci_configure_ltr(struct pci_dev *dev)
{
#ifdef CONFIG_PCIEASPM
2016
	struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
2017 2018 2019
	u32 cap;
	struct pci_dev *bridge;

2020 2021 2022
	if (!host->native_ltr)
		return;

2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
	if (!pci_is_pcie(dev))
		return;

	pcie_capability_read_dword(dev, PCI_EXP_DEVCAP2, &cap);
	if (!(cap & PCI_EXP_DEVCAP2_LTR))
		return;

	/*
	 * Software must not enable LTR in an Endpoint unless the Root
	 * Complex and all intermediate Switches indicate support for LTR.
	 * PCIe r3.1, sec 6.18.
	 */
	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT)
		dev->ltr_path = 1;
	else {
		bridge = pci_upstream_bridge(dev);
		if (bridge && bridge->ltr_path)
			dev->ltr_path = 1;
	}

	if (dev->ltr_path)
		pcie_capability_set_word(dev, PCI_EXP_DEVCTL2,
					 PCI_EXP_DEVCTL2_LTR_EN);
#endif
}

2049 2050 2051 2052 2053
static void pci_configure_device(struct pci_dev *dev)
{
	struct hotplug_params hpp;
	int ret;

2054
	pci_configure_mps(dev);
2055
	pci_configure_extended_tags(dev, NULL);
2056
	pci_configure_relaxed_ordering(dev);
2057
	pci_configure_ltr(dev);
2058

2059 2060 2061 2062 2063 2064 2065 2066 2067 2068
	memset(&hpp, 0, sizeof(hpp));
	ret = pci_get_hp_params(dev, &hpp);
	if (ret)
		return;

	program_hpp_type2(dev, hpp.t2);
	program_hpp_type1(dev, hpp.t1);
	program_hpp_type0(dev, hpp.t0);
}

2069 2070 2071
static void pci_release_capabilities(struct pci_dev *dev)
{
	pci_vpd_release(dev);
2072
	pci_iov_release(dev);
2073
	pci_free_cap_save_buffers(dev);
2074 2075
}

L
Linus Torvalds 已提交
2076
/**
B
Bjorn Helgaas 已提交
2077 2078
 * pci_release_dev - Free a PCI device structure when all users of it are
 *		     finished
L
Linus Torvalds 已提交
2079 2080
 * @dev: device that's been disconnected
 *
B
Bjorn Helgaas 已提交
2081
 * Will be called only by the device core when all users of this PCI device are
L
Linus Torvalds 已提交
2082 2083 2084 2085
 * done.
 */
static void pci_release_dev(struct device *dev)
{
2086
	struct pci_dev *pci_dev;
L
Linus Torvalds 已提交
2087

2088
	pci_dev = to_pci_dev(dev);
2089
	pci_release_capabilities(pci_dev);
2090
	pci_release_of_node(pci_dev);
2091
	pcibios_release_device(pci_dev);
2092
	pci_bus_put(pci_dev->bus);
2093
	kfree(pci_dev->driver_override);
2094
	kfree(pci_dev->dma_alias_mask);
L
Linus Torvalds 已提交
2095 2096 2097
	kfree(pci_dev);
}

2098
struct pci_dev *pci_alloc_dev(struct pci_bus *bus)
2099 2100 2101 2102 2103 2104 2105 2106
{
	struct pci_dev *dev;

	dev = kzalloc(sizeof(struct pci_dev), GFP_KERNEL);
	if (!dev)
		return NULL;

	INIT_LIST_HEAD(&dev->bus_list);
2107
	dev->dev.type = &pci_dev_type;
2108
	dev->bus = pci_bus_get(bus);
2109 2110 2111

	return dev;
}
2112 2113
EXPORT_SYMBOL(pci_alloc_dev);

2114 2115 2116 2117 2118
static bool pci_bus_crs_vendor_id(u32 l)
{
	return (l & 0xffff) == 0x0001;
}

S
Sinan Kaya 已提交
2119 2120
static bool pci_bus_wait_crs(struct pci_bus *bus, int devfn, u32 *l,
			     int timeout)
L
Linus Torvalds 已提交
2121 2122 2123
{
	int delay = 1;

S
Sinan Kaya 已提交
2124 2125
	if (!pci_bus_crs_vendor_id(*l))
		return true;	/* not a CRS completion */
L
Linus Torvalds 已提交
2126

S
Sinan Kaya 已提交
2127 2128
	if (!timeout)
		return false;	/* CRS, but caller doesn't want to wait */
L
Linus Torvalds 已提交
2129

2130
	/*
S
Sinan Kaya 已提交
2131 2132 2133
	 * We got the reserved Vendor ID that indicates a completion with
	 * Configuration Request Retry Status (CRS).  Retry until we get a
	 * valid Vendor ID or we time out.
2134
	 */
2135
	while (pci_bus_crs_vendor_id(*l)) {
S
Sinan Kaya 已提交
2136
		if (delay > timeout) {
2137 2138 2139 2140
			pr_warn("pci %04x:%02x:%02x.%d: not ready after %dms; giving up\n",
				pci_domain_nr(bus), bus->number,
				PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);

2141
			return false;
L
Linus Torvalds 已提交
2142
		}
2143 2144 2145 2146
		if (delay >= 1000)
			pr_info("pci %04x:%02x:%02x.%d: not ready after %dms; waiting\n",
				pci_domain_nr(bus), bus->number,
				PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);
2147

L
Linus Torvalds 已提交
2148 2149
		msleep(delay);
		delay *= 2;
2150

2151 2152
		if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
			return false;
L
Linus Torvalds 已提交
2153 2154
	}

2155 2156 2157 2158 2159
	if (delay >= 1000)
		pr_info("pci %04x:%02x:%02x.%d: ready after %dms\n",
			pci_domain_nr(bus), bus->number,
			PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);

2160 2161
	return true;
}
S
Sinan Kaya 已提交
2162

2163 2164
bool pci_bus_generic_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
					int timeout)
S
Sinan Kaya 已提交
2165 2166 2167 2168
{
	if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
		return false;

B
Bjorn Helgaas 已提交
2169
	/* Some broken boards return 0 or ~0 if a slot is empty: */
S
Sinan Kaya 已提交
2170 2171 2172 2173 2174 2175 2176
	if (*l == 0xffffffff || *l == 0x00000000 ||
	    *l == 0x0000ffff || *l == 0xffff0000)
		return false;

	if (pci_bus_crs_vendor_id(*l))
		return pci_bus_wait_crs(bus, devfn, l, timeout);

2177 2178
	return true;
}
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196

bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
				int timeout)
{
#ifdef CONFIG_PCI_QUIRKS
	struct pci_dev *bridge = bus->self;

	/*
	 * Certain IDT switches have an issue where they improperly trigger
	 * ACS Source Validation errors on completions for config reads.
	 */
	if (bridge && bridge->vendor == PCI_VENDOR_ID_IDT &&
	    bridge->device == 0x80b5)
		return pci_idt_bus_quirk(bus, devfn, l, timeout);
#endif

	return pci_bus_generic_read_dev_vendor_id(bus, devfn, l, timeout);
}
2197 2198 2199
EXPORT_SYMBOL(pci_bus_read_dev_vendor_id);

/*
B
Bjorn Helgaas 已提交
2200 2201
 * Read the config data for a PCI device, sanity-check it,
 * and fill in the dev structure.
2202 2203 2204 2205 2206 2207 2208 2209 2210
 */
static struct pci_dev *pci_scan_device(struct pci_bus *bus, int devfn)
{
	struct pci_dev *dev;
	u32 l;

	if (!pci_bus_read_dev_vendor_id(bus, devfn, &l, 60*1000))
		return NULL;

2211
	dev = pci_alloc_dev(bus);
L
Linus Torvalds 已提交
2212 2213 2214 2215 2216 2217
	if (!dev)
		return NULL;

	dev->devfn = devfn;
	dev->vendor = l & 0xffff;
	dev->device = (l >> 16) & 0xffff;
2218

2219 2220
	pci_set_of_node(dev);

Y
Yu Zhao 已提交
2221
	if (pci_setup_device(dev)) {
2222
		pci_bus_put(dev->bus);
L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228 2229
		kfree(dev);
		return NULL;
	}

	return dev;
}

2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
static void pcie_report_downtraining(struct pci_dev *dev)
{
	if (!pci_is_pcie(dev))
		return;

	/* Look from the device up to avoid downstream ports with no devices */
	if ((pci_pcie_type(dev) != PCI_EXP_TYPE_ENDPOINT) &&
	    (pci_pcie_type(dev) != PCI_EXP_TYPE_LEG_END) &&
	    (pci_pcie_type(dev) != PCI_EXP_TYPE_UPSTREAM))
		return;

	/* Multi-function PCIe devices share the same link/status */
	if (PCI_FUNC(dev->devfn) != 0 || dev->is_virtfn)
		return;

	/* Print link status only if the device is constrained by the fabric */
	__pcie_print_link_status(dev, false);
}

2249 2250
static void pci_init_capabilities(struct pci_dev *dev)
{
2251 2252 2253
	/* Enhanced Allocation */
	pci_ea_init(dev);

2254 2255
	/* Setup MSI caps & disable MSI/MSI-X interrupts */
	pci_msi_setup_pci_dev(dev);
2256

2257 2258 2259
	/* Buffers for saving PCIe and PCI-X capabilities */
	pci_allocate_cap_save_buffers(dev);

2260 2261 2262 2263
	/* Power Management */
	pci_pm_init(dev);

	/* Vital Product Data */
2264
	pci_vpd_init(dev);
Y
Yu Zhao 已提交
2265 2266

	/* Alternative Routing-ID Forwarding */
2267
	pci_configure_ari(dev);
2268 2269 2270

	/* Single Root I/O Virtualization */
	pci_iov_init(dev);
2271

2272 2273 2274
	/* Address Translation Services */
	pci_ats_init(dev);

2275
	/* Enable ACS P2P upstream forwarding */
C
Chris Wright 已提交
2276
	pci_enable_acs(dev);
2277

2278 2279
	/* Precision Time Measurement */
	pci_ptm_init(dev);
B
Bjorn Helgaas 已提交
2280

K
Keith Busch 已提交
2281 2282
	/* Advanced Error Reporting */
	pci_aer_init(dev);
2283

2284 2285
	pcie_report_downtraining(dev);

2286 2287
	if (pci_probe_reset_function(dev) == 0)
		dev->reset_fn = 1;
2288 2289
}

M
Marc Zyngier 已提交
2290
/*
B
Bjorn Helgaas 已提交
2291
 * This is the equivalent of pci_host_bridge_msi_domain() that acts on
M
Marc Zyngier 已提交
2292 2293 2294 2295 2296 2297 2298 2299
 * devices. Firmware interfaces that can select the MSI domain on a
 * per-device basis should be called from here.
 */
static struct irq_domain *pci_dev_msi_domain(struct pci_dev *dev)
{
	struct irq_domain *d;

	/*
B
Bjorn Helgaas 已提交
2300
	 * If a domain has been set through the pcibios_add_device()
M
Marc Zyngier 已提交
2301 2302 2303 2304 2305 2306
	 * callback, then this is the one (platform code knows best).
	 */
	d = dev_get_msi_domain(&dev->dev);
	if (d)
		return d;

2307 2308 2309 2310 2311 2312 2313 2314
	/*
	 * Let's see if we have a firmware interface able to provide
	 * the domain.
	 */
	d = pci_msi_get_device_domain(dev);
	if (d)
		return d;

M
Marc Zyngier 已提交
2315 2316 2317
	return NULL;
}

2318 2319
static void pci_set_msi_domain(struct pci_dev *dev)
{
M
Marc Zyngier 已提交
2320 2321
	struct irq_domain *d;

2322
	/*
M
Marc Zyngier 已提交
2323 2324 2325
	 * If the platform or firmware interfaces cannot supply a
	 * device-specific MSI domain, then inherit the default domain
	 * from the host bridge itself.
2326
	 */
M
Marc Zyngier 已提交
2327 2328 2329 2330 2331
	d = pci_dev_msi_domain(dev);
	if (!d)
		d = dev_get_msi_domain(&dev->bus->dev);

	dev_set_msi_domain(&dev->dev, d);
2332 2333
}

2334
void pci_device_add(struct pci_dev *dev, struct pci_bus *bus)
L
Linus Torvalds 已提交
2335
{
2336 2337
	int ret;

2338 2339
	pci_configure_device(dev);

2340 2341
	device_initialize(&dev->dev);
	dev->dev.release = pci_release_dev;
L
Linus Torvalds 已提交
2342

2343
	set_dev_node(&dev->dev, pcibus_to_node(bus));
2344
	dev->dev.dma_mask = &dev->dma_mask;
2345
	dev->dev.dma_parms = &dev->dma_parms;
2346
	dev->dev.coherent_dma_mask = 0xffffffffull;
L
Linus Torvalds 已提交
2347

2348
	pci_set_dma_max_seg_size(dev, 65536);
2349
	pci_set_dma_seg_boundary(dev, 0xffffffff);
2350

L
Linus Torvalds 已提交
2351 2352 2353
	/* Fix up broken headers */
	pci_fixup_device(pci_fixup_header, dev);

B
Bjorn Helgaas 已提交
2354
	/* Moved out from quirk header fixup code */
2355 2356
	pci_reassigndev_resource_alignment(dev);

B
Bjorn Helgaas 已提交
2357
	/* Clear the state_saved flag */
2358 2359
	dev->state_saved = false;

2360 2361
	/* Initialize various capabilities */
	pci_init_capabilities(dev);
2362

L
Linus Torvalds 已提交
2363 2364 2365 2366
	/*
	 * Add the device to our list of discovered devices
	 * and the bus list for fixup functions, etc.
	 */
2367
	down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
2368
	list_add_tail(&dev->bus_list, &bus->devices);
2369
	up_write(&pci_bus_sem);
2370 2371 2372 2373

	ret = pcibios_add_device(dev);
	WARN_ON(ret < 0);

B
Bjorn Helgaas 已提交
2374
	/* Set up MSI IRQ domain */
2375 2376
	pci_set_msi_domain(dev);

2377 2378 2379 2380
	/* Notifier could use PCI capabilities */
	dev->match_driver = false;
	ret = device_add(&dev->dev);
	WARN_ON(ret < 0);
2381 2382
}

2383
struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn)
2384 2385 2386
{
	struct pci_dev *dev;

T
Trent Piepho 已提交
2387 2388 2389 2390 2391 2392
	dev = pci_get_slot(bus, devfn);
	if (dev) {
		pci_dev_put(dev);
		return dev;
	}

2393 2394 2395 2396 2397
	dev = pci_scan_device(bus, devfn);
	if (!dev)
		return NULL;

	pci_device_add(dev, bus);
L
Linus Torvalds 已提交
2398 2399 2400

	return dev;
}
2401
EXPORT_SYMBOL(pci_scan_single_device);
L
Linus Torvalds 已提交
2402

2403
static unsigned next_fn(struct pci_bus *bus, struct pci_dev *dev, unsigned fn)
M
Matthew Wilcox 已提交
2404
{
2405 2406 2407
	int pos;
	u16 cap = 0;
	unsigned next_fn;
2408

2409 2410 2411 2412 2413 2414
	if (pci_ari_enabled(bus)) {
		if (!dev)
			return 0;
		pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI);
		if (!pos)
			return 0;
2415

2416 2417 2418 2419
		pci_read_config_word(dev, pos + PCI_ARI_CAP, &cap);
		next_fn = PCI_ARI_CAP_NFN(cap);
		if (next_fn <= fn)
			return 0;	/* protect against malformed list */
M
Matthew Wilcox 已提交
2420

2421 2422 2423 2424 2425 2426
		return next_fn;
	}

	/* dev may be NULL for non-contiguous multifunction devices */
	if (!dev || dev->multifunction)
		return (fn + 1) % 8;
M
Matthew Wilcox 已提交
2427 2428 2429 2430 2431 2432

	return 0;
}

static int only_one_child(struct pci_bus *bus)
{
2433
	struct pci_dev *bridge = bus->self;
2434

2435 2436 2437 2438 2439
	/*
	 * Systems with unusual topologies set PCI_SCAN_ALL_PCIE_DEVS so
	 * we scan for all possible devices, not just Device 0.
	 */
	if (pci_has_flag(PCI_SCAN_ALL_PCIE_DEVS))
M
Matthew Wilcox 已提交
2440
		return 0;
2441 2442

	/*
2443 2444 2445 2446 2447 2448 2449
	 * A PCIe Downstream Port normally leads to a Link with only Device
	 * 0 on it (PCIe spec r3.1, sec 7.3.1).  As an optimization, scan
	 * only for Device 0 in that situation.
	 *
	 * Checking has_secondary_link is a hack to identify Downstream
	 * Ports because sometimes Switches are configured such that the
	 * PCIe Port Type labels are backwards.
2450
	 */
2451
	if (bridge && pci_is_pcie(bridge) && bridge->has_secondary_link)
M
Matthew Wilcox 已提交
2452
		return 1;
2453

M
Matthew Wilcox 已提交
2454 2455 2456
	return 0;
}

L
Linus Torvalds 已提交
2457
/**
B
Bjorn Helgaas 已提交
2458
 * pci_scan_slot - Scan a PCI slot on a bus for devices
L
Linus Torvalds 已提交
2459
 * @bus: PCI bus to scan
B
Bjorn Helgaas 已提交
2460
 * @devfn: slot number to scan (must have zero function)
L
Linus Torvalds 已提交
2461 2462 2463
 *
 * Scan a PCI slot on the specified PCI bus for devices, adding
 * discovered devices to the @bus->devices list.  New devices
2464
 * will not have is_added set.
2465 2466
 *
 * Returns the number of new devices found.
L
Linus Torvalds 已提交
2467
 */
2468
int pci_scan_slot(struct pci_bus *bus, int devfn)
L
Linus Torvalds 已提交
2469
{
M
Matthew Wilcox 已提交
2470
	unsigned fn, nr = 0;
2471
	struct pci_dev *dev;
M
Matthew Wilcox 已提交
2472 2473 2474

	if (only_one_child(bus) && (devfn > 0))
		return 0; /* Already scanned the entire slot */
L
Linus Torvalds 已提交
2475

2476
	dev = pci_scan_single_device(bus, devfn);
2477 2478 2479
	if (!dev)
		return 0;
	if (!dev->is_added)
2480 2481
		nr++;

2482
	for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) {
M
Matthew Wilcox 已提交
2483 2484 2485 2486 2487
		dev = pci_scan_single_device(bus, devfn + fn);
		if (dev) {
			if (!dev->is_added)
				nr++;
			dev->multifunction = 1;
L
Linus Torvalds 已提交
2488 2489
		}
	}
S
Shaohua Li 已提交
2490

B
Bjorn Helgaas 已提交
2491
	/* Only one slot has PCIe device */
2492
	if (bus->self && nr)
S
Shaohua Li 已提交
2493 2494
		pcie_aspm_init_link_state(bus->self);

L
Linus Torvalds 已提交
2495 2496
	return nr;
}
2497
EXPORT_SYMBOL(pci_scan_slot);
L
Linus Torvalds 已提交
2498

2499 2500 2501 2502 2503 2504 2505
static int pcie_find_smpss(struct pci_dev *dev, void *data)
{
	u8 *smpss = data;

	if (!pci_is_pcie(dev))
		return 0;

2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519
	/*
	 * We don't have a way to change MPS settings on devices that have
	 * drivers attached.  A hot-added device might support only the minimum
	 * MPS setting (MPS=128).  Therefore, if the fabric contains a bridge
	 * where devices may be hot-added, we limit the fabric MPS to 128 so
	 * hot-added devices will work correctly.
	 *
	 * However, if we hot-add a device to a slot directly below a Root
	 * Port, it's impossible for there to be other existing devices below
	 * the port.  We don't limit the MPS in this case because we can
	 * reconfigure MPS on both the Root Port and the hot-added device,
	 * and there are no other devices involved.
	 *
	 * Note that this PCIE_BUS_SAFE path assumes no peer-to-peer DMA.
2520
	 */
2521 2522
	if (dev->is_hotplug_bridge &&
	    pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT)
2523 2524 2525 2526 2527 2528 2529 2530 2531 2532
		*smpss = 0;

	if (*smpss > dev->pcie_mpss)
		*smpss = dev->pcie_mpss;

	return 0;
}

static void pcie_write_mps(struct pci_dev *dev, int mps)
{
2533
	int rc;
2534 2535

	if (pcie_bus_config == PCIE_BUS_PERFORMANCE) {
2536
		mps = 128 << dev->pcie_mpss;
2537

2538 2539
		if (pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT &&
		    dev->bus->self)
B
Bjorn Helgaas 已提交
2540 2541 2542

			/*
			 * For "Performance", the assumption is made that
2543 2544 2545 2546 2547
			 * downstream communication will never be larger than
			 * the MRRS.  So, the MPS only needs to be configured
			 * for the upstream communication.  This being the case,
			 * walk from the top down and set the MPS of the child
			 * to that of the parent bus.
2548 2549 2550 2551 2552
			 *
			 * Configure the device MPS with the smaller of the
			 * device MPSS or the bridge MPS (which is assumed to be
			 * properly configured at this point to the largest
			 * allowable MPS based on its parent bus).
2553
			 */
2554
			mps = min(mps, pcie_get_mps(dev->bus->self));
2555 2556 2557 2558
	}

	rc = pcie_set_mps(dev, mps);
	if (rc)
2559
		pci_err(dev, "Failed attempting to set the MPS\n");
2560 2561
}

2562
static void pcie_write_mrrs(struct pci_dev *dev)
2563
{
2564
	int rc, mrrs;
2565

B
Bjorn Helgaas 已提交
2566 2567
	/*
	 * In the "safe" case, do not configure the MRRS.  There appear to be
2568 2569 2570 2571 2572
	 * issues with setting MRRS to 0 on a number of devices.
	 */
	if (pcie_bus_config != PCIE_BUS_PERFORMANCE)
		return;

B
Bjorn Helgaas 已提交
2573 2574
	/*
	 * For max performance, the MRRS must be set to the largest supported
2575
	 * value.  However, it cannot be configured larger than the MPS the
2576
	 * device or the bus can support.  This should already be properly
B
Bjorn Helgaas 已提交
2577
	 * configured by a prior call to pcie_write_mps().
2578
	 */
2579
	mrrs = pcie_get_mps(dev);
2580

B
Bjorn Helgaas 已提交
2581 2582
	/*
	 * MRRS is a R/W register.  Invalid values can be written, but a
2583
	 * subsequent read will verify if the value is acceptable or not.
2584 2585
	 * If the MRRS value provided is not acceptable (e.g., too large),
	 * shrink the value until it is acceptable to the HW.
2586
	 */
2587 2588
	while (mrrs != pcie_get_readrq(dev) && mrrs >= 128) {
		rc = pcie_set_readrq(dev, mrrs);
2589 2590
		if (!rc)
			break;
2591

2592
		pci_warn(dev, "Failed attempting to set the MRRS\n");
2593 2594
		mrrs /= 2;
	}
2595 2596

	if (mrrs < 128)
2597
		pci_err(dev, "MRRS was unable to be configured with a safe value.  If problems are experienced, try running with pci=pcie_bus_safe\n");
2598 2599 2600 2601
}

static int pcie_bus_configure_set(struct pci_dev *dev, void *data)
{
J
Jon Mason 已提交
2602
	int mps, orig_mps;
2603 2604 2605 2606

	if (!pci_is_pcie(dev))
		return 0;

2607 2608
	if (pcie_bus_config == PCIE_BUS_TUNE_OFF ||
	    pcie_bus_config == PCIE_BUS_DEFAULT)
2609 2610
		return 0;

J
Jon Mason 已提交
2611 2612
	mps = 128 << *(u8 *)data;
	orig_mps = pcie_get_mps(dev);
2613 2614

	pcie_write_mps(dev, mps);
2615
	pcie_write_mrrs(dev);
2616

2617
	pci_info(dev, "Max Payload Size set to %4d/%4d (was %4d), Max Read Rq %4d\n",
2618
		 pcie_get_mps(dev), 128 << dev->pcie_mpss,
J
Jon Mason 已提交
2619
		 orig_mps, pcie_get_readrq(dev));
2620 2621 2622 2623

	return 0;
}

B
Bjorn Helgaas 已提交
2624 2625
/*
 * pcie_bus_configure_settings() requires that pci_walk_bus work in a top-down,
2626 2627 2628
 * parents then children fashion.  If this changes, then this code will not
 * work as designed.
 */
2629
void pcie_bus_configure_settings(struct pci_bus *bus)
2630
{
2631
	u8 smpss = 0;
2632

2633
	if (!bus->self)
2634 2635 2636
		return;

	if (!pci_is_pcie(bus->self))
2637 2638
		return;

B
Bjorn Helgaas 已提交
2639 2640
	/*
	 * FIXME - Peer to peer DMA is possible, though the endpoint would need
2641
	 * to be aware of the MPS of the destination.  To work around this,
2642 2643 2644 2645 2646
	 * simply force the MPS of the entire system to the smallest possible.
	 */
	if (pcie_bus_config == PCIE_BUS_PEER2PEER)
		smpss = 0;

2647
	if (pcie_bus_config == PCIE_BUS_SAFE) {
2648
		smpss = bus->self->pcie_mpss;
2649

2650 2651 2652 2653 2654 2655 2656
		pcie_find_smpss(bus->self, &smpss);
		pci_walk_bus(bus, pcie_find_smpss, &smpss);
	}

	pcie_bus_configure_set(bus->self, &smpss);
	pci_walk_bus(bus, pcie_bus_configure_set, &smpss);
}
2657
EXPORT_SYMBOL_GPL(pcie_bus_configure_settings);
2658

2659 2660 2661 2662 2663 2664 2665 2666 2667
/*
 * Called after each bus is probed, but before its children are examined.  This
 * is marked as __weak because multiple architectures define it.
 */
void __weak pcibios_fixup_bus(struct pci_bus *bus)
{
       /* nothing to do, expected to be removed in the future */
}

2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681
/**
 * pci_scan_child_bus_extend() - Scan devices below a bus
 * @bus: Bus to scan for devices
 * @available_buses: Total number of buses available (%0 does not try to
 *		     extend beyond the minimal)
 *
 * Scans devices below @bus including subordinate buses. Returns new
 * subordinate number including all the found devices. Passing
 * @available_buses causes the remaining bus space to be distributed
 * equally between hotplug-capable bridges to allow future extension of the
 * hierarchy.
 */
static unsigned int pci_scan_child_bus_extend(struct pci_bus *bus,
					      unsigned int available_buses)
L
Linus Torvalds 已提交
2682
{
2683 2684
	unsigned int used_buses, normal_bridges = 0, hotplug_bridges = 0;
	unsigned int start = bus->busn_res.start;
2685
	unsigned int devfn, fn, cmax, max = start;
L
Linus Torvalds 已提交
2686
	struct pci_dev *dev;
2687
	int nr_devs;
L
Linus Torvalds 已提交
2688

B
Bjorn Helgaas 已提交
2689
	dev_dbg(&bus->dev, "scanning bus\n");
L
Linus Torvalds 已提交
2690 2691

	/* Go find them, Rover! */
2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707
	for (devfn = 0; devfn < 256; devfn += 8) {
		nr_devs = pci_scan_slot(bus, devfn);

		/*
		 * The Jailhouse hypervisor may pass individual functions of a
		 * multi-function device to a guest without passing function 0.
		 * Look for them as well.
		 */
		if (jailhouse_paravirt() && nr_devs == 0) {
			for (fn = 1; fn < 8; fn++) {
				dev = pci_scan_single_device(bus, devfn + fn);
				if (dev)
					dev->multifunction = 1;
			}
		}
	}
L
Linus Torvalds 已提交
2708

B
Bjorn Helgaas 已提交
2709
	/* Reserve buses for SR-IOV capability */
2710 2711
	used_buses = pci_iov_bus_range(bus);
	max += used_buses;
2712

L
Linus Torvalds 已提交
2713 2714 2715 2716
	/*
	 * After performing arch-dependent fixup of the bus, look behind
	 * all PCI-to-PCI bridges on this bus.
	 */
A
Alex Chiang 已提交
2717
	if (!bus->is_added) {
B
Bjorn Helgaas 已提交
2718
		dev_dbg(&bus->dev, "fixups for bus\n");
A
Alex Chiang 已提交
2719
		pcibios_fixup_bus(bus);
2720
		bus->is_added = 1;
A
Alex Chiang 已提交
2721 2722
	}

2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734
	/*
	 * Calculate how many hotplug bridges and normal bridges there
	 * are on this bus. We will distribute the additional available
	 * buses between hotplug bridges.
	 */
	for_each_pci_bridge(dev, bus) {
		if (dev->is_hotplug_bridge)
			hotplug_bridges++;
		else
			normal_bridges++;
	}

2735 2736 2737 2738 2739
	/*
	 * Scan bridges that are already configured. We don't touch them
	 * unless they are misconfigured (which will be done in the second
	 * scan below).
	 */
2740 2741 2742
	for_each_pci_bridge(dev, bus) {
		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, max, 0, 0);
2743 2744 2745 2746 2747 2748 2749 2750

		/*
		 * Reserve one bus for each bridge now to avoid extending
		 * hotplug bridges too much during the second scan below.
		 */
		used_buses++;
		if (cmax - max > 1)
			used_buses += cmax - max - 1;
2751
	}
2752 2753

	/* Scan bridges that need to be reconfigured */
2754 2755 2756 2757
	for_each_pci_bridge(dev, bus) {
		unsigned int buses = 0;

		if (!hotplug_bridges && normal_bridges == 1) {
B
Bjorn Helgaas 已提交
2758

2759 2760 2761 2762 2763 2764 2765 2766
			/*
			 * There is only one bridge on the bus (upstream
			 * port) so it gets all available buses which it
			 * can then distribute to the possible hotplug
			 * bridges below.
			 */
			buses = available_buses;
		} else if (dev->is_hotplug_bridge) {
B
Bjorn Helgaas 已提交
2767

2768 2769 2770 2771 2772
			/*
			 * Distribute the extra buses between hotplug
			 * bridges if any.
			 */
			buses = available_buses / hotplug_bridges;
2773
			buses = min(buses, available_buses - used_buses + 1);
2774 2775 2776 2777
		}

		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, cmax, buses, 1);
2778 2779 2780
		/* One bus is already accounted so don't add it again */
		if (max - cmax > 1)
			used_buses += max - cmax - 1;
2781
	}
L
Linus Torvalds 已提交
2782

2783 2784
	/*
	 * Make sure a hotplug bridge has at least the minimum requested
2785 2786
	 * number of buses but allow it to grow up to the maximum available
	 * bus number of there is room.
2787
	 */
2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
	if (bus->self && bus->self->is_hotplug_bridge) {
		used_buses = max_t(unsigned int, available_buses,
				   pci_hotplug_bus_size - 1);
		if (max - start < used_buses) {
			max = start + used_buses;

			/* Do not allocate more buses than we have room left */
			if (max > bus->busn_res.end)
				max = bus->busn_res.end;

			dev_dbg(&bus->dev, "%pR extended by %#02x\n",
				&bus->busn_res, max - start);
		}
2801 2802
	}

L
Linus Torvalds 已提交
2803 2804 2805 2806 2807 2808 2809
	/*
	 * We've scanned the bus and so we know all about what's on
	 * the other side of any bridges that may be on this bus plus
	 * any devices.
	 *
	 * Return how far we've got finding sub-buses.
	 */
B
Bjorn Helgaas 已提交
2810
	dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max);
L
Linus Torvalds 已提交
2811 2812
	return max;
}
2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824

/**
 * pci_scan_child_bus() - Scan devices below a bus
 * @bus: Bus to scan for devices
 *
 * Scans devices below @bus including subordinate buses. Returns new
 * subordinate number including all the found devices.
 */
unsigned int pci_scan_child_bus(struct pci_bus *bus)
{
	return pci_scan_child_bus_extend(bus, 0);
}
2825
EXPORT_SYMBOL_GPL(pci_scan_child_bus);
L
Linus Torvalds 已提交
2826

2827
/**
B
Bjorn Helgaas 已提交
2828 2829
 * pcibios_root_bridge_prepare - Platform-specific host bridge setup
 * @bridge: Host bridge to set up
2830 2831 2832 2833 2834 2835 2836 2837 2838
 *
 * Default empty implementation.  Replace with an architecture-specific setup
 * routine, if necessary.
 */
int __weak pcibios_root_bridge_prepare(struct pci_host_bridge *bridge)
{
	return 0;
}

2839 2840 2841 2842 2843 2844 2845 2846
void __weak pcibios_add_bus(struct pci_bus *bus)
{
}

void __weak pcibios_remove_bus(struct pci_bus *bus)
{
}

2847 2848
struct pci_bus *pci_create_root_bus(struct device *parent, int bus,
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
L
Linus Torvalds 已提交
2849
{
2850
	int error;
2851
	struct pci_host_bridge *bridge;
L
Linus Torvalds 已提交
2852

2853
	bridge = pci_alloc_host_bridge(0);
2854
	if (!bridge)
2855
		return NULL;
2856 2857

	bridge->dev.parent = parent;
2858

2859 2860 2861 2862
	list_splice_init(resources, &bridge->windows);
	bridge->sysdata = sysdata;
	bridge->busnr = bus;
	bridge->ops = ops;
2863

2864 2865 2866
	error = pci_register_host_bridge(bridge);
	if (error < 0)
		goto err_out;
2867

2868
	return bridge->bus;
L
Linus Torvalds 已提交
2869 2870

err_out:
2871
	kfree(bridge);
L
Linus Torvalds 已提交
2872 2873
	return NULL;
}
2874
EXPORT_SYMBOL_GPL(pci_create_root_bus);
2875

2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908
int pci_host_probe(struct pci_host_bridge *bridge)
{
	struct pci_bus *bus, *child;
	int ret;

	ret = pci_scan_root_bus_bridge(bridge);
	if (ret < 0) {
		dev_err(bridge->dev.parent, "Scanning root bridge failed");
		return ret;
	}

	bus = bridge->bus;

	/*
	 * We insert PCI resources into the iomem_resource and
	 * ioport_resource trees in either pci_bus_claim_resources()
	 * or pci_bus_assign_resources().
	 */
	if (pci_has_flag(PCI_PROBE_ONLY)) {
		pci_bus_claim_resources(bus);
	} else {
		pci_bus_size_bridges(bus);
		pci_bus_assign_resources(bus);

		list_for_each_entry(child, &bus->children, node)
			pcie_bus_configure_settings(child);
	}

	pci_bus_add_devices(bus);
	return 0;
}
EXPORT_SYMBOL_GPL(pci_host_probe);

2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924
int pci_bus_insert_busn_res(struct pci_bus *b, int bus, int bus_max)
{
	struct resource *res = &b->busn_res;
	struct resource *parent_res, *conflict;

	res->start = bus;
	res->end = bus_max;
	res->flags = IORESOURCE_BUS;

	if (!pci_is_root_bus(b))
		parent_res = &b->parent->busn_res;
	else {
		parent_res = get_pci_domain_busn_res(pci_domain_nr(b));
		res->flags |= IORESOURCE_PCI_FIXED;
	}

2925
	conflict = request_resource_conflict(parent_res, res);
2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971

	if (conflict)
		dev_printk(KERN_DEBUG, &b->dev,
			   "busn_res: can not insert %pR under %s%pR (conflicts with %s %pR)\n",
			    res, pci_is_root_bus(b) ? "domain " : "",
			    parent_res, conflict->name, conflict);

	return conflict == NULL;
}

int pci_bus_update_busn_res_end(struct pci_bus *b, int bus_max)
{
	struct resource *res = &b->busn_res;
	struct resource old_res = *res;
	resource_size_t size;
	int ret;

	if (res->start > bus_max)
		return -EINVAL;

	size = bus_max - res->start + 1;
	ret = adjust_resource(res, res->start, size);
	dev_printk(KERN_DEBUG, &b->dev,
			"busn_res: %pR end %s updated to %02x\n",
			&old_res, ret ? "can not be" : "is", bus_max);

	if (!ret && !res->parent)
		pci_bus_insert_busn_res(b, res->start, res->end);

	return ret;
}

void pci_bus_release_busn_res(struct pci_bus *b)
{
	struct resource *res = &b->busn_res;
	int ret;

	if (!res->flags || !res->parent)
		return;

	ret = release_resource(res);
	dev_printk(KERN_DEBUG, &b->dev,
			"busn_res: %pR %s released\n",
			res, ret ? "can not be" : "is");
}

2972
int pci_scan_root_bus_bridge(struct pci_host_bridge *bridge)
2973
{
2974
	struct resource_entry *window;
2975
	bool found = false;
2976
	struct pci_bus *b;
2977
	int max, bus, ret;
2978

2979 2980 2981 2982
	if (!bridge)
		return -EINVAL;

	resource_list_for_each_entry(window, &bridge->windows)
2983 2984 2985 2986
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
2987

2988 2989 2990 2991 2992 2993
	ret = pci_register_host_bridge(bridge);
	if (ret < 0)
		return ret;

	b = bridge->bus;
	bus = bridge->busnr;
2994

2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006
	if (!found) {
		dev_info(&b->dev,
		 "No busn resource found for root bus, will use [bus %02x-ff]\n",
			bus);
		pci_bus_insert_busn_res(b, bus, 255);
	}

	max = pci_scan_child_bus(b);

	if (!found)
		pci_bus_update_busn_res_end(b, max);

3007
	return 0;
3008
}
3009
EXPORT_SYMBOL(pci_scan_root_bus_bridge);
3010 3011 3012 3013

struct pci_bus *pci_scan_root_bus(struct device *parent, int bus,
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
{
3014
	struct resource_entry *window;
3015
	bool found = false;
3016
	struct pci_bus *b;
3017 3018
	int max;

3019
	resource_list_for_each_entry(window, resources)
3020 3021 3022 3023
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
3024

3025
	b = pci_create_root_bus(parent, bus, ops, sysdata, resources);
3026 3027 3028
	if (!b)
		return NULL;

3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040
	if (!found) {
		dev_info(&b->dev,
		 "No busn resource found for root bus, will use [bus %02x-ff]\n",
			bus);
		pci_bus_insert_busn_res(b, bus, 255);
	}

	max = pci_scan_child_bus(b);

	if (!found)
		pci_bus_update_busn_res_end(b, max);

3041
	return b;
3042
}
3043 3044
EXPORT_SYMBOL(pci_scan_root_bus);

B
Bill Pemberton 已提交
3045
struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
3046 3047 3048 3049 3050 3051 3052
					void *sysdata)
{
	LIST_HEAD(resources);
	struct pci_bus *b;

	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
3053
	pci_add_resource(&resources, &busn_resource);
3054 3055
	b = pci_create_root_bus(NULL, bus, ops, sysdata, &resources);
	if (b) {
3056
		pci_scan_child_bus(b);
3057 3058 3059 3060 3061 3062 3063
	} else {
		pci_free_resource_list(&resources);
	}
	return b;
}
EXPORT_SYMBOL(pci_scan_bus);

3064
/**
B
Bjorn Helgaas 已提交
3065
 * pci_rescan_bus_bridge_resize - Scan a PCI bus for devices
3066 3067 3068 3069 3070 3071 3072 3073 3074
 * @bridge: PCI bridge for the bus to scan
 *
 * Scan a PCI bus and child buses for new devices, add them,
 * and enable them, resizing bridge mmio/io resource if necessary
 * and possible.  The caller must ensure the child devices are already
 * removed for resizing to occur.
 *
 * Returns the max number of subordinate bus discovered.
 */
3075
unsigned int pci_rescan_bus_bridge_resize(struct pci_dev *bridge)
3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088
{
	unsigned int max;
	struct pci_bus *bus = bridge->subordinate;

	max = pci_scan_child_bus(bus);

	pci_assign_unassigned_bridge_resources(bridge);

	pci_bus_add_devices(bus);

	return max;
}

3089
/**
B
Bjorn Helgaas 已提交
3090
 * pci_rescan_bus - Scan a PCI bus for devices
3091 3092
 * @bus: PCI bus to scan
 *
B
Bjorn Helgaas 已提交
3093 3094
 * Scan a PCI bus and child buses for new devices, add them,
 * and enable them.
3095 3096 3097
 *
 * Returns the max number of subordinate bus discovered.
 */
3098
unsigned int pci_rescan_bus(struct pci_bus *bus)
3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109
{
	unsigned int max;

	max = pci_scan_child_bus(bus);
	pci_assign_unassigned_bus_resources(bus);
	pci_bus_add_devices(bus);

	return max;
}
EXPORT_SYMBOL_GPL(pci_rescan_bus);

3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127
/*
 * pci_rescan_bus(), pci_rescan_bus_bridge_resize() and PCI device removal
 * routines should always be executed under this mutex.
 */
static DEFINE_MUTEX(pci_rescan_remove_lock);

void pci_lock_rescan_remove(void)
{
	mutex_lock(&pci_rescan_remove_lock);
}
EXPORT_SYMBOL_GPL(pci_lock_rescan_remove);

void pci_unlock_rescan_remove(void)
{
	mutex_unlock(&pci_rescan_remove_lock);
}
EXPORT_SYMBOL_GPL(pci_unlock_rescan_remove);

R
Ryan Desfosses 已提交
3128 3129
static int __init pci_sort_bf_cmp(const struct device *d_a,
				  const struct device *d_b)
3130
{
3131 3132 3133
	const struct pci_dev *a = to_pci_dev(d_a);
	const struct pci_dev *b = to_pci_dev(d_b);

3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145
	if      (pci_domain_nr(a->bus) < pci_domain_nr(b->bus)) return -1;
	else if (pci_domain_nr(a->bus) > pci_domain_nr(b->bus)) return  1;

	if      (a->bus->number < b->bus->number) return -1;
	else if (a->bus->number > b->bus->number) return  1;

	if      (a->devfn < b->devfn) return -1;
	else if (a->devfn > b->devfn) return  1;

	return 0;
}

3146
void __init pci_sort_breadthfirst(void)
3147
{
3148
	bus_sort_breadthfirst(&pci_bus_type, &pci_sort_bf_cmp);
3149
}
3150 3151 3152 3153

int pci_hp_add_bridge(struct pci_dev *dev)
{
	struct pci_bus *parent = dev->bus;
3154
	int busnr, start = parent->busn_res.start;
3155
	unsigned int available_buses = 0;
3156 3157 3158 3159 3160 3161 3162
	int end = parent->busn_res.end;

	for (busnr = start; busnr <= end; busnr++) {
		if (!pci_find_bus(pci_domain_nr(parent), busnr))
			break;
	}
	if (busnr-- > end) {
3163
		pci_err(dev, "No bus number available for hot-added bridge\n");
3164 3165
		return -1;
	}
3166 3167 3168 3169

	/* Scan bridges that are already configured */
	busnr = pci_scan_bridge(parent, dev, busnr, 0);

3170 3171 3172 3173 3174 3175
	/*
	 * Distribute the available bus numbers between hotplug-capable
	 * bridges to make extending the chain later possible.
	 */
	available_buses = end - busnr;

3176
	/* Scan bridges that need to be reconfigured */
3177
	pci_scan_bridge_extend(parent, dev, busnr, available_buses, 1);
3178

3179 3180 3181 3182 3183 3184
	if (!dev->subordinate)
		return -1;

	return 0;
}
EXPORT_SYMBOL_GPL(pci_hp_add_bridge);